Home

UM0492 User manual

1. Rotor _ Angle void STO_Calc_Rotor_Angle Volt_Components Stat_Volt_alfa_beta Curr_Components Stat_Curr_alfa_beta s16 hBusVoltage It is the core of the module as it implements the State observer this function has to be called with the same periodicity of stator current sampling in the demo program since that periodicity coincides with the FOC execution rate as discussed in Section 4 2 it is called from inside the FOC routine It gets the measured stator currents Stat_Curr_alfa_beta the applied voltage commands Stat_Volt_alfa_beta and the measured DC bus voltage hBusVoltage as inputs at step k as a result it carries out step k 1 of the discretized state observer equations thus achieving estimation of the motor back emf e and ep Consequently by means of a numerical PLL back emfs are processed to calculate rotor speed and angle Observed back emfs observed rotor angle and speed are written into module private variables Stator voltage commands vand Vp s16 format measured stator currents i and ig s16 format DC bus voltage s16 format See MC_type h for structure declarations None See Section 2 2 for more information about the sensorless algorithm ky UM0492 Library functions STO Calc Speed Synopsis Description Input Returns Note void STO_Calc_Speed void This function has to be called with the timing of the speed loop control in the demo program it is fixed by th
2. UM0492 Y User manual STM32F103xx permanent magnet synchronous motor FOC software library V1 0 January 2008 Introduction This user manual describes the permanent magnet synchronous motor PMSM FOC software library a field oriented control FOC firmware library for 3 phase permanent magnet motors developed for the STM32F103xx microcontrollers These 32 bit ARM Cortex M3 cored ST microcontrollers STM32F103xx come with a set of peripherals that makes it suitable for performing both permanent magnet and AC induction motor FOC In particular this manual describes the STM32F103xx software library developed to control sine wave driven permanent magnet motors in both torque and speed control mode These motors may be equipped with an encoder with three Hall sensors or they may be sensorless The control of an AC induction motor equipped with encoder or tacho generator is described in the UM0483 user manual The PMSM FOC is made of several C modules compatible with the free of charge IAR EWARM KickStart edition toolchain version 4 42 It is used to quickly evaluate both the MCU and the available tools In addition when used together with the STM32F103xx motor control starter kit STM3210B MCKIT and PM motor a motor can be made to run in a very short time It also eliminates the need for time consuming development of FOC and speed regulation algorithms by providing ready to use functions that let the user concentrate on
3. For this analysis the motor electrical characteristics are assumed to be isotropic with respect to the q and d axes So it is assumed that the torque and flux regulators have the same starting value of Kp and that they also have the same K value 113 121 Additional information UM0492 Figure 58 shows the closed loop system in which the motor phase is modelled using the resistor inductance equivalent circuit in the locked rotor condition Block A is the proportionality constant between the software variable storing the voltage command expressed in digit and the real voltage applied to the motor phase expressed in Volt Likewise block B is the is the proportionality constant between the real current expressed in Ampere and the software variable storing the phase current expressed in digit Figure 58 Closed loop block diagram Voltage command Target current Measured current ai14851 The transfer functions of the two blocks A and B are expressed by the following formulas Vande RshuntAop2 Bu 2 A and B SC respectively By putting Kp K L Rg it is possible to perform pole zero cancellation as described in Figure 59 Figure 59 Pole zero cancellation Voltage K command Target current Measured current ai14852 In this condition the closed loop system is brought back to a first order system and the dyna
4. where 9 is the angular displacement of the q d reference frame at the time of observation and 8 that displacement at t 0 see Figure 45 Figure 45 Transformation from an abc stationary frame to a rotating frame q d ai14844 With Clark s transformation stator currents ias and ips which are directed along axes each displaced by 120 degrees are resolved into currents i and i on a stationary reference frame B 75 121 Library functions UM0492 5 4 3 76 121 Appropriate substitution into the general equations given above yields la las _ last Zins ig T In Park s change of variables stator currents i and ig which belong to a stationary reference frame B are resolved to a reference frame synchronous with the rotor and oriented so that the d axis is aligned with the permanent magnets flux so as to obtain igs and igs Consequently with this choice of reference we have igs c0SO igsin igs gS NO igcosO On the other hand reverse Park transformation takes back stator voltage vg and vg belonging to a rotating frame synchronous and properly oriented with the rotor to a stationary reference frame so as to obtain va and vg Va Vgg C088 VgsS NO Vg VggSiN Vgs COSO Circle limitation As discussed above FOC allows to separately control the torque and the flux of a 3 phase permanent magnet motor After the two new values Vy and V
5. 2 T A Lipo and D W Novotny Vector Control and Dynamics of AC Drives Oxford University Press 1996 e 3 S Morimoto Y Takeda T Hirasa K Taniguchi Expansion of Operating Limits for Permanent Magnet Motor by Optimum Flux Weakening Conference Record of the 1989 IEEE pp 51 56 1989 119 121 Revision history UM0492 Revision history 120 121 Table 5 Document revision history Date Revision Changes 08 Jan 2008 1 Initial release UM0492 Please Read Carefully Information in this document is provided solely in connection with ST products STMicroelectronics NV and its subsidiaries ST reserve the right to make changes corrections modifications or improvements to this document and the products and services described herein at any time without notice All ST products are sold pursuant to ST s terms and conditions of sale Purchasers are solely responsible for the choice selection and use of the ST products and services described herein and ST assumes no liability whatsoever relating to the choice selection or use of the ST products and services described herein No license express or implied by estoppel or otherwise to any intellectual property rights is granted under this document If any part of this document refers to any third party products or services it shall not be deemed a license grant by ST for the use of such third party products or services or any intel
6. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 DBG_I2C 2_SMBU Reserved S _TIMEO UT Res rw 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 DBG_I2C 1_SMBU DBG_ DBG_ DBG_ DBG_ DBG_ DBG_ DBG_ TRACE_ TRACE_ DBG_ DBG_ DBG_ S TIME CAN_ TIM4_ TIM3_ TIM2_ TIMi_ WWDG_ IWDG MODE IOEN Reserved STANDB STOP SLEEP OU STOP STOP STOP STOP STOP STOP STOP 1 0 kd rw rw rw rw rw rw rw rw rw rw rw Res rw rw rw 38 121 Ti UM0492 Getting started with the library 4 4 1 Getting started with the library It is quite easy to set up an operational evaluation platform with a drive system that includes the STM3210B MCKIT featuring the STM32F103xx microcontroller where this software runs and a permanent magnet motor This section explains how to quickly configure your system and if necessary customize the library accordingly Follow these steps to accomplish this task 1 Gather all the information that is needed regarding the hardware in use motor parameters power devices features speed position sensor parameters current sensors transconductance 2 Edit using an IDE the stm32 10x_MCconf h configuration header file as explained in more detail in Section 4 1 and the following parameter header files MC_Control_Param h see Section 4 2 MC_encoder_param h see Section 4 3 or MC_hall_prm h see Section 4 4 MC_State_Observer_param h see Section 4 5 MC_
7. RCC_ADCCLKConfig RCC_AHBPeriphClockCmd RCC_APB2PeriphClockCmd GPIO_Structlnit GPIO_Init GPIO_PinLockConfig GPIO_PinRemapConfig TIM1_Delnit TIM1_TimeBaseStructinit TIM1_TimeBaselnit TIM1_OCStructinit TIM1_OC1 Init TIM1_OC2Init TIM1_OC3init TIM1_BDTRConfig TIM1_SelectOutputTrigger TIM1_ClearlTPendingBit TIM1_ITConfig TIM1_Cmd ADC_Delnit ADC_Cmd ADC_Structlnit ADC_Init ADC_StartCalibration ADC_GetCalibrationStatus ADC_InjectedSequencerLengthConfig ADC_InjectedChannelContfig ADC_ExternalTriginjectedConvCmd NVIC_PriorityGroupConfig NVIC_Structlnit NVIC_Init Motor control library SVPWM_lIcsCurrentReadingCalibration 65 121 Library functions UM0492 66 121 SVPWM_lIcsCurrentReadingCalibration Synopsis Description Input Returns Note Functions called void SVPWM_IcsCurrentReadingCalibration void The purpose of this function is to store the two analog voltages corresponding to zero current values for compensating the offset introduced by both ICS and amplification network None None This function is called by MCL_Init which is executed at every motor startup It reads the analog voltage on the A D channels used for current reading before the PWM outputs are enabled so that the current flowing through the inverter is zero Standard Library ADC_ITConfig ADC_ExternalTrigInjectedConvConfig ADC_ExternalTrigInjectedConvCmd ADC_InjectedChannelConfig ADC_ClearFlag ADC_SoftwareS
8. Functions called void TIM1_UP_IRQHandler void This interrupt handler is executed after an update event when an underflow of the TIM1 counter occurs It is used only in three shunt configuration to externally trigger the ADC when a rising edge of the TIM1_OC4Ref signal occurs None None ADC_ClearFlag TIM1_ClearFlag UM0492 Library functions ADC_IRQHandler Synopsis Description Input Returns Functions called See also void ADC_IRQHandler void The purpose of this function is to handle the ADC global interrupt request Two different possible interrupt sources are managed JEOC end of conversion injected group AWD analog watchdog JEOC if the main state machine is in the Start state it triggers the motor startup procedure which depends upon the system configuration see Section 4 1 otherwise if the state is Run it triggers the execution of the FOC algorithm If the system configuration includes the brake resistor see Section 3 12 it manages its hysteresis switching in case of overvoltage If the DAC functionality is enabled see Section 3 5 it updates the value of the variables of interest AWD in the event of an overvoltage it switches on the brake resistor or generates a fault OVER_VOLTAGE depending if BRAKE _RESISTOR is commented in stm32f10x_MCconf h see Section 4 1 and Section 3 12 None None FOC_Model ADC_GetITStatus ADC_ClearFlag MCL_Calc_Bus
9. Ty 2 55 us Tsg 0 7 us TrR 2 55 us The maximum applicable duty cycles are listed in Table 2 as a function of the PWM frequency The values in Table 2 are measured using the MB459 board This evaluation platform is designed to support several motor driving topologies PMSM and AC induction and current reading strategies single and three shunt resistor Therefore the figures provided in Table 2 should be understood as a starting point and not as a best case Table 2 PWM frequency vs maximum duty cycle relationship PWM frequency Max duty cycle Max modulation Index Up to 11 4 kHz 100 100 12 2 kHz 99 98 12 9 kHz 98 5 97 13 7 kHz 98 96 14 4 kHz 98 96 15 2 kHz 97 94 16 kHz 96 5 93 16 7 kHz 96 5 93 17 5 kHz 95 5 91 It is possible to adjust the noise parameters based on customized hardware by editing the following definitions in the MC_pwm_3shunt_prm h header file define SAMPLING_TIME_NS 700 0 7usec define TNOISE_NS 2550 2 55usec define TRISE_NS 2550 2 55usec Changing the noise parameters sampling time and dead time affects the values provided in Table 2 Isolated current sensor reading and space vector PWM generation stm32 10x_svpwm_ics module Two important tasks are performed in the stm32 10x_svpwm_ics module Space vector pulse width modulation SVPWM Three phase current reading when two isolated current sensors ICS are
10. zr r isthe stator phase winding resistance Li isthe stator phase winding leakage inductance Lysis the stator phase winding magnetizing inductance 40 is the rotor electrical angle is the flux linkage due to permanent magnets The complexity of these equations is apparent as the three stator flux linkages are mutually coupled and what is more as they are dependent on the rotor position which is time varying and a function of the electromagnetic and load torques The reference frame theory simplifies the PM motor equations by making a change of variables that refers the stator quantities abc that can be visualized as directed along axes each 120 apart to qd components directed along a 90 apart axes rotating synchronously with the rotor and vice versa see Section 5 4 for more details The d direct axis is aligned with the rotor flux while the q quadrature axis leads at 90 degrees in the positive rolling direction The motor voltage and flux equations are simplified to da Lsig da Leig O a Va Teig TE t oa where Ls Lis 3 2 Lms da Va lsa y Aa The electromagnetic torque equation becomes 3 3 a P a 5P Ag lq dala 5P Leiig ig Brig Leig ig 3 Te 5P Png The last equation makes it clear that the quadrature current component igs has linear control on the torque generation whereas the current component ig has no effect on it as
11. of the stator voltage producing flux and torque components of the stator current have been independently comput Sa flux and torque PIDs it is necessary to saturate the magnitude of the resulting vector ita before passing them to the Reverse Park transformation and finally to the SVPWM block The saturation boundary is normally given by the value S16_MAX 32767 which produces the maximum output voltage magnitude corresponding to a duty cycle going from 0 to 100 Nevertheless when using three shunt resistor configuration and depending on PWM frequency it might be necessary to limit the maximum PWM duty cycle to guarantee the proper functioning of the stator currents reading block For this reason the saturation boundary could be a value slightly lower than S16_ MAX depending on PWM switching frequency when using three shunt resistor configuration Table 2 on page 64 repeated below for convenience shows the maximum applicable modulation index as a function of the PWM switching frequency when using the STM3210B MCKIT UM0492 Library functions Table 3 PWM frequency vs maximum duty cycle relationship PWM frequency Max duty cycle Max modulation Index Up to 11 4 kHz 100 100 12 2 kHz 99 98 12 9 kHz 98 5 97 13 7 kHz 98 96 14 4 kHz 98 96 15 2 kHz 97 94 16 kHz 96 5 93 16 7 kHz 96 5 93 17 5 kHz 95 5 91 Note The figures above were measured using the MB459 board
12. and Mc_State_Observer Interface modules The Mc_State_Observer module designed for permanent magnet synchronous motors implements a back emf state observer and a phase locked loop PLL It is able to detect rotor angular position and speed In addition the module processes the output data and by doing so implements a safety feature that detects locked rotor condition or malfunctioning The MC_State_Observer_Interface module acts as an interface with the first providing motor parameters and state observer default gains The MC_State_Observer module which is the engine of the sensorless algorithm is provided as a compiled object file the source code is available free of charge from ST on request please contact your nearest ST sales office UM0492 Library functions 5 7 1 List of available C functions The following is a list of available functions as listed in the MC_State_Observer_Interface h and MC_State_Observer h header files STO_Init on page 89 STO_StateObserverlnterface_Init on page 90 STO_Obs_Gains_Update on page 90 STO_Calc_Rotor_Angle on page 90 STO_Calc_Speed on page 91 STO_InitSpeedBuffer on page 91 STO_Get_Electrical_Angle on page 91 STO_Get_Mechanical_Angle on page 91 STO_Get_Speed on page 92 STO_Get_Speed_Hz on page 92 STO_IsSpeed_Reliable on page 92 STO_Check_Speed_Reliability on page 92 STO_Start_Up on page 93 STO_Get_wlalfa_est on page 93 STO_Get_wlibeta_est on page 93 STO_Get_wBemf_alfa_est on p
13. corresponding to the offset must be subtracted when reading phase current A D converted values Therefore the function must be called after SVPWM_3ShuntCurrentReadingCalibration None 55 121 Library functions UM0492 56 121 SVPWM_3ShuntCalcDutyCycles Synopsis Description Input Returns Note void SVPWM_3ShuntCalcDutyCycles Volt_Components Stat_Volt_Input After executing the FOC algorithm new stator voltage components Va and Vg are computed The purpose of this function is to calculate exactly the three duty cycles to be applied to the inverter legs starting from the values of these voltage components Moreover once the three duty cycles to be applied in the next PWM period are known this function sets the Channel 4 output compare register used to set the sampling point for the next current reading In particular depending on the duty cycle values the sampling point is computed see Section 5 1 3 Refer to Section 5 1 2 for information on the theoretical approach of SVPWM Va and Vg None None Functions called None SVPWM_3ShuntAdvCurrentReading Synopsis Description Input Returns Note Functions called void SVPWM_3ShuntAdvCurrentReading FunctionalState cmd It is used to enable or disable advanced current reading If advanced current reading is disabled current reading is performed in conjunction with the update event cmd ENABLE or DISABLE None None TIM1_ClearFla
14. plus 11 7 Kbytes for LCD joystick management 9 121 Getting started with tools UM0492 1 1 1 2 10 121 Getting started with tools To develop an application for a PM synchronous motor using the PMSM FOC software library you must set up a complete development environment as described in the following sections A PC running Windows XP is necessary Working environment The PMSM FOC software library was fully validated using the main hardware boards included in STM3210B MCKIT starter kit a complete inverter and control board The STM3210B MCKIT starter kit provides an ideal toolset for starting a project and using the library Therefore for rapid implementation and evaluation of the software described in this user manual it is recommended to acquire this starter kit It is also recommended to install the IAR EWARM C toolchain version 4 42 which was used to compile the PMSM FOC software library With this toolchain you do not need to configure your workspace You can set up your workspace manually for any other toolchain A free kickstart edition of the IAR EWARM C toolchain with a 32Kb limitation can be downloaded from www iar com it is sufficient to compile and evaluate the software library presented here Software tools A complete software package consists of eA third party integrated development environment IDE eA third party C compiler This library was compiled using the third party IAR C toolchain e JT
15. speed feedback which is based on the state observer theory A state observer in control theory is a system that provides an estimation of the internal state of a real system given its input and output measurement In our case the internal states of the motor are the back emfs and the phase currents while the input and output quantities supplied are the phase voltages and measured currents respectively see Figure 9 DC bus voltage measurement is used to convert voltage commands into voltage applied to motor phases UM0492 Introduction to the sensorless FOC of PM motors Figure 9 General sensorless algorithm block diagram Power converter m e 157 ghunts laps In particular the observed states are compared for consistency with the real system via the phase currents and the result is used to adjust the model through a gain vector K1 K2 The motor back emfs are defined as ea PnP cos pat ep pA sin pat As can be seen they hold information about the rotor angle Then back emfs are fed to a block which acting as a PLL is able to reconstruct the rotor electrical angle and speed Figure 10 shows a scope capture taken while the motor is running in field oriented control positive rolling direction the yellow and the red waveforms C1 C2 are respectively the observed back emfs alpha and beta the blue square wave C3 is a signal coming from a Hall sensor c
16. the magnets are glued to the surface of the rotor and this is the reason why it is referred to as SM PMSM surface mounted PMSM b in the second one illustrated by drawings b and c in Figure 7 the magnets are embedded in the rotor structure This construction is known as IPMSM interior PMSM Figure 7 Different PM motor constructions d I I aq Ta q a b c SM PMSMs inherently have an isotropic structure that is the direct and quadrature inductances Lg and L are the same Usually their mechanical structure allows a wider airgap which in turn means lower flux weakening capability thus setting a limit to the maximum speed On the other hand IPMSMs show an anisotropic structure slight in the b construction called inset PM motor strong in the c configuration called buried or radial PM motor their fine mechanical structure usually shows a narrow airgap thus giving good flux weakening capability This firmware library is optimized for use in conjunction with SM PMSMs and to some extent with inset PM machines UM0492 Introduction to the sensorless FOC of PM motors 2 1 2 PMSM field oriented control FOC fundamental equations With reference to Figure 8 the motor voltage and flux linkage equations of an SM PMSM are generally expressed as Spi abc Vabc slabo dt L L Lis Lms L Loo sino z 2r Aabo ms Lis Lms 3s labo r 3 where 2T L L sin
17. 0 2 00 e eee eee 26 LCD screen for setting the P term of flux PID 0 0 00 ccc eee eee 26 LCD screen for setting the P term of the speed PID 00 020 cea euee 26 LCD screen for setting the P term of the flux PID 0 0 0 0 c eee eee 27 LCD screen for setting the P term of the flux PID 0 0000 c eee eee 27 Power stage StalUS eeso rrepu bee de ke aes Rees eee doe ee ie Mae aoe 28 Error message shown in the event of an undervoltage fault 0000 29 ICS hardware connections 0 000 31 Encoder output signals counter operation 000 00 e eee 32 60 and 120 displaced Hall sensor output waveforms 00 0 eee ee eee 33 Determination of Hall electrical phase shift 0 00 ee 34 Brake resistor Circuit 0000 ccc eee 37 DBG_TIM1_STOP bit in TIM1 control register extract from STM32 reference manual 38 Alignment angle aii eee tee cee pe kee ENTS alice oly E EERE EEE EREE Manel 45 Startup current system frequency and amplitude profile 000 0000s 49 Voa and VB stator voltage components 0 cee eens 57 SVPWM phase voltages waveforms 00 c cee 57 PWM and ADC synchronization 00000 eee eee 59 Inverter leg and shunt resistor position 2 2 0 0 cee eee 60 Low side switches gate signals low modulation indexes 20055 61 Low side Phase A duty cycle gt DT TN 0 0
18. 121 Current regulators fine tuning To fine tune the current regulator it is required to start with the parameters K and Kp computed following the instruction of appendix A 4 A priori determination of flux and torque current PI gains Then starting from the default configuration of stm32f10x_MCconf h follow the following steps Fill the power devices parameters current regulation parameters power board protections thresholds and speed loop sampling time sections of MC_Control_ Param h as described in Section 4 2 e Instm32f10x_MCconf h select the kind of sensor to be used during the development stage of your design and fill the related header file MC_encoder_param h or MC_hall_param h as described in Section 4 3 or Section 4 4 The tuning of the current regulators is not supported in sensorless configuration Fill in NOMINAL_CURRENT and POLE_PAIRS in MC_PMSM_motor_param h e Uncomment FLUX_TORQUE_PIDs_TUNING A firmware generating a square wave shaped reference torque will be generated The firmware generated when FLUX_TORQUE_PIDs_TUNING is not commented must be run only in Torque contro mode A square wave amplitude and period can be selected by editing PID_TORQUE_REFERENCE and SQUARE_WAVE_PERIOD parameters in MC_Control_Param h The goal is to tune the torque and flux current components PIDs in real time For this purpose t
19. 121 Disabling the linear curve computation routine stm32 10x_Timebase module If you want to disable the linear curve computation you must comment out the PID_Speed_Coefficients_update routine In this case the default values for Ki Kp Kd for torque flux and speed regulation are used See PID_TORQUE_Kx_DEFAULT PID_FLUX_Kx_DEFAULT PID_SPEED_Kx_DEFAULT in the MC_control_Param h file To disable the linear curve computation routine in stm32 10x_Timebase c void SysTickHandler void J if wGlobal_Flags amp SPEED CONTROL SPEED CONTROL if State RUN PID_Speed_Coefficients_update XXX_Get_Speed to be commented General purpose time base stm32 10x_Timebase module The purpose of the stm32 10x_Timebase module is to generate a time base that can be used by the other modules of the applications List of available functions and interrupt service routines The following is a list of available functions as listed in the stm32 10x_Timebase h header file e 7B _Init on page 101 e TB Waiton page 101 e 7B _StartUp_Timeout_lsElapsed TB_Delay_IsElapsed TB_DisplayDelay_IsElapsed TB_DebounceDelay_IsElapsed on page 102 e TB _Set_Delay_500us TB_Set_DisplayDelay_500us TB_Set_StartUp_ Timeout TB_Set_DebounceDelay_500us on page 101 e SysTickHandler on page 102 UM0492 Library functions TB Init Synopsis Description Input Retur
20. 5 Encoder feedback processing stm32 10x_encoder module 78 5 5 1 List of available functions and interrupt service routines 78 5 6 Hall sensor feedback processing stm32 10x_hall module 81 5 6 1 List of available functions asas aaa aaa aaa 81 5 6 2 Speed measurement implementation a n anaana 85 5 6 3 Electrical angle extrapolation implementation 87 5 7 Sensorless speed position detection MC_State_Observer and MC_State_Observer_Interface modules 88 5 7 1 List of available C functions 0 000 c eee eee 89 5 8 Currents and speed regulators MC_PID_regulators module 94 5 8 1 List of available functions 0 0 0 eee eee 94 5 8 2 PID regulator theoretical background 00 eee eee eee 97 5 8 3 Regulator sampling time setting 0 eee eee 97 5 8 4 Adjusting speed regulation loop Ki Kp and Kd vs motor frequency 98 5 9 General purpose time base stm32 10x_Timebase module 100 5 9 1 List of available functions and interrupt service routines 100 5 10 Power stage check up MC_MotorControl_Layer module 103 5 10 1 List of available functions 0 2 0 0 00 103 5 11 Main interrupt service routines stm32 10x_it module 107 5 11 1 List of non empty interrupt service routines 108 Appendix A Additional information 0000 cece eee eee eee 110 A 1 Adjus
21. ISR changing repetition counter has a direct impact on FOC refresh rate and thus on CPU load However in the case of three shunt topology current reading to ensure that the FOC algorithm is executed once for each PWM register update it is necessary to keep the synchronization between current conversions triggering and PWM register update In the proposed software library this is automatically performed so that you can reduce the frequency of execution of the FOC algorithm by simply changing the default value of the repetition counter the REP_RATE parameter in the MC_Control_Param h header file Figure 55 shows current sampling triggering and FOC algorithm execution with respect to PWM period when REP_RATE is set to 3 Figure 55 AD conversions for three shunt topology stator currents reading and FOC algorithm execution when REP_RATE 3 and PWM frequency gt 18 kHz PWM Period REP_RATE 3 PWM frequency gt 18 kHz E ADC triggering interval E FOC algorithm execution New duty cycle values ready ai14839c Because three shunt resistor topology requires low side switches to be on when performing current reading A D conversions the REP_RATE parameter must be an odd number in this case Considering that the raw FOC algorithm execution time is about 24 us in sensorless and three shunt resistor current reading configuration the related contribution to CPU load can be computed as follows _____Fewm REP_RATE 1 2
22. Ina which for the proposed hardware can be estimated approximately at Ima 0 6 Repunt SO the per unit current value is obtained by dividing the physical value by that base UM0492 Additional information A 4 In this way ip is always in the range from 1 to 1 Therefore the q1 15 format which ranges from 1 to 0 999969482421 875 with a resolution of 0 00003051 7578125 is perfectly suitable taking care of the overflow value 1 1 1 and thus extensively used Thus the complete transformation equation from SI units is i S l 5 QF igt 15 floor 7 2 max A priori determination of flux and torque current PI gains The aim of this appendix is to provide a criterion for the computation of the initial values of the torque flux PI parameters K and Kp Appendix A 5 discusses the way of fine tuning them To calculate these starting values it is required to know the electrical characteristics of the motor stator resistance R and inductance L and the electrical characteristics of the hardware shunt resistor Rgpunt current sense amplification network Ao and the direct current bus voltage VgusDC The derivative action of the controller is not considered using this method Figure 57 shows the PI controller block diagram used for torque or flux regulation Figure 57 Block diagram of PI controller Target Voltage command current Measured current ai14850
23. The DAC functionality was implemented in the presented firmware library by using two out of the four TIM3 output compare channels PBO and PB1 pins and by modulating the duty cycle of the generated 30 kHz PWM signal In order to properly filter the generated signals without introducing important delays on the waveforms it is suggested to use a proper first order low pass filter e g with a 10 kQ resistor and a 22 nF capacitor Power stage feedbacks A dedicated menu was designed to show the value in volts of the DC bus voltage and the temperature of the STM3210B MCKIT power board heat sink Figure 22 Power stage status STM32 Motor Control PMSM FOC ver 1 0 Power Stage Status DC bus 024 Volt 020 Celsius lt Move Fault messages This section provides a description of all the possible fault messages that can be detected when using the software library together with the STM3210B MCKIT Figure 23 shows a typical error message as displayed on the LCD UM0492 Running the demo program 3 7 1 3 7 2 3 7 3 Note Figure 23 Error message shown in the event of an undervoltage fault STM32 Motor Control PMSM FOC ver 1 90 fFAULT Bus Under Voltage Press Key to return to menu The message Press Key to return to menu is visible only if the source of the fault has disappeared In this case pressing the Key button causes the main state machine to switc
24. and flux regulators by the FOC_CalcFluxTorqueRef function while in Torque control mode they are settled by the user Therefore as a current source is required the function has to run the power converter as a CR PWM For this purpose it implements an high performance synchronous d q frame current regulator whose operating frequency is defined as explained in Section 4 2 by the parameter REP_RATE in conjunction with PWM_FREQ Triggered by ADC JEOC ISR the function loads stator currents read by ICS or shunt resistors and carries out Clark and Park transformations converting them to ig and igs See Figure 5 Then these currents are fed to PID regulators together with reference values ig and igs The regulator output voltages vg and Vgs then must be transformed back to a stator frame through Reverse Park conversion and finally drive the power stage In order to correctly perform Park and Reverse Park transformation it is essential to know the rotor position 0 9 because currents have to be oriented in phase and in quadrature with the rotor flux To manage this task depending on the configuration set in stm32 10x_MCconf h the function can read the rotor angle measurement from encoders Hall sensors or the provided sensorless algorithm None None Clarke Park RevPark_Circle_Limitation PID_Torque_Regulator PID_Flux_Regulator Rev_Park If working with encoder ENC_Get_Electrical_Angle if Working
25. cece eae 62 DT Ty Ts 2 lt ADuty a lt D TN and ADuty 4g lt Dr Tp Ts ee seer ees ak a AS ald Se ta 62 ADutyA lt DT Tyt Tg 2 and ADutya p gt DT Tptts 1 0 0 cee eee 63 ADutya lt DT Tyt Tg 2 and ADutya p lt DT Tptls 00 eee 63 Stator currents sampling in ICS configuration REP_RATE 1 0 0000eeuee 68 Clarke Park and reverse Park transformations 0 0 00 cece eee eee eae 71 Radians versus 16 ieran eaei ea a ia aua eee eee 74 s16 versus sine and cosine 1 0 fetes 74 Transformation from an abc stationary frame to a rotating frame q d 75 Circle limitation working principle 0 aaea 77 Hall sensors output state correspondence nannaa 85 Hall sensor timer interface prescaler decrease n nnna ananuna 86 7 121 List of figures UM0492 Figure 49 Figure 50 Figure 51 Figure 52 Figure 53 Figure 54 Figure 55 Figure 56 Figure 57 Figure 58 Figure 59 Figure 60 Figure 61 Figure 62 8 121 Hall sensor timer interface prescaler increase 00 000 eee 86 TIMx_IRQHandler flowchart 0 0 00 eee eee eee eens 87 Hall sensor output transitions 0 2 eae 88 PID general equation 0 eee 97 Time domain to discrete PID equations 0 cece eee 98 Linear curve for coefficient computation 2 2 0 cee eee 99 AD conversions for three shunt topology stator currents reading and FOC algorithm ex
26. default the final phase A current value can be computed by using the formula Alignment final phase A current I_LALIGNMENT 0 64 32767 Rehunt ky 45 121 Getting started with the library UM0492 4 4 46 121 Hall sensor parameters MC_hall_prm h The MC_hall_prm h header file is to be filled if a position speed sensing is performed by means of three Hall sensors HALL_SENSORS or VIEW_HALL_FEEDBACK defined in stm32 10x_MCconf h define TIMER2_ HANDLES HALL To be uncommented if the three sensor output signals are wired to TIMER2 input pins default required if using STM3210B MCKIT define TIMER3_HANDLES HALL To be uncommented if the three sensor output signals are wired to TIMERS input pins define TIMER4_ HANDLES HALL To be uncommented if the three sensor output signals are wired to TIMER4 input pins define HALL _SENSORS_PLACEMENT Defines the electrical displacement between the Hall sensors expressed in degrees physical displacement x number of pole pairs The choice is between 120 DEGREES_120 and 60 degrees DEGREES_60 define HALL_PHASE_SHIFT Defines the electrical phase shift degrees between the 0 angle with the convention utilized in the firmware and the first rising edge on TIMx_CH1 H1 signal with STM3210B MCKIT Refer to Section 3 10 for a detailed explanation on how to determine this parameter define HALL MAX SPEED _FDBK_RPM Defines the rotor mechani
27. distinguished Parameters needed to carry out the FOC e Parameters needed to perform sensorless rotor speed position detection e Parameters needed to perform the flux weakening functionality Basic motor parameters This is the minimum set of motor parameters to be known in order to carry out FOC define POLE_PAIR_NUM Defines the number of magnetic pole pairs define NOMINAL_CURRENT In digits defines the motor nominal current 0 to peak according to the formula I Amps Rehunt A 65536 digi LAMPS Penunt Ay 65596 i l Of course these data are to be matched with inverter current rating UM0492 Getting started with the library Note 4 6 2 4 6 3 define MOTOR_RATED_SPEED_RPM In RPM defines the motor rated speed This parameter could be set depending on the mechanical construction of the motor the application specifications or it could be worked out using the provided spreadsheet see Section 4 6 3 If the flux weakening functionality is not to be used then define EXT_SPEED_RANGE should be set to 1 Motor parameters for sensorless FOC These parameters are used by the state observer algorithm to detect the rotor speed and position see also Section 4 5 define RS Defines the motor winding resistance phase in Ohms define LS Defines the motor winding inductance phase in Henry define MOTOR_VOLTAGE_CONSTANT Defines the motor
28. frequency 2 REP_RATE 3 Fixed point numerical representation The PMSM FOC software library uses fixed point representation of fractional signed values Thus a number nis expressed as n m f where mis the integer part magnitude and f the fractional part and both m and fhave fixed numbers of digits In terms of two s complement binary representation if a variable n requires QI bits to express as powers of two its magnitude of which 1 bit is needed for the sign QF bits as inverse powers of two for its fractional part then we have to allocate QI QF bits for that variable Therefore given a choice of QI and QF the variable representation has the following features e Range 2 Q1 1 lt n lt 2 Ql 1 2 QPF e Resolution 1 29F The equation below converts a fractional quantity q to fixed point representation n n floor q 2 A common way to express the choice that has been made is the q QI QF notation So if a variable is stored in q3 5 format it means that 3 bits are reserved for the magnitude 5 bits for the resolution the expressible range is from 4 to 3 96875 the resolution is 0 03125 the bit weighting is bit n 7 6 5 4 3 2 1 0 value 4 2 1 1 2 1 4 1 8 1 16 1 32 This software library uses the PU Per Unit system to express current values They are always referred to a base quantity that is the maximum measurable current
29. gains by following the instructions given in appendix A 4 and A 5 Comment FLUX_TORQUE_PIDs_TUNING run the motor in speed control and tune the speed PI D gains Derivative action can be enabled disabled by uncommenting commenting the SPEED_DIFFERENTIAL_ TERM ENABLED line in stm32f10x_MCcont h 2 Runthe motor in sensor mode and tune the observer gains The Clark and Parke transformation blocks and the speed regulator will utilize the rotor position and speed information read by the sensor The sensorless algorithm will be run in parallel To this purpose it is necessary to Precompute initial observer gains following the instructions provided in appendix A 6 Uncomment OBSERVER_GAIN_TUNING in stm32f10x_MCcont h Fillin MC_PMSM_motor_param h see Section 4 6 and MC_State_Observer_param h startup section not required in this step see Section 4 5 Real time tune observer gains by visually comparing through DAC functionality the observed and real and Ip error must be null and by making sure that the observed B emf waveforms are as clean and sinusoidal as possible Real time tuning of PLL gains should not be required but in case proportionally increase Kp and Ki to enlarge the PLL bandwidth quickness to speed variation and vice versa Once the 4 gains have been found write them in MC_State_Observer_param h 35 121 Running the demo progra
30. mentioned above these equations are valid for SM PMSMs 17 121 Introduction to the sensorless FOC of PM motors UM0492 2 2 18 121 Therefore if is the motor rated current then its maximum torque is produced for igs ls and igs 0 in fact I Jig ig On the other hand the magnetic flux can be weakened by acting on the direct axis current igs this extends the achievable speed range but at the cost of a decrease in quadrature current igs and hence in the electromagnetic torque supplied to the load see Section 2 3 for details about the Flux weakening strategy In conclusion by regulating the motor currents through their components ig and igs FOC manages to regulate the PMSM torque and flux current regulation is achieved by means of what is usually called a synchronous frame CR PWM Figure 8 Assumed PMSM reference frame convention Introduction to sensorless rotor position speed feedback In Section 2 1 it has been shown that rotor position speed measurement has a crucial role in PMSM field oriented control Hall sensors or encoders are broadly used in the control chain for that purpose Sensorless algorithms for rotor position speed feedback are considered very useful and for different reasons to lower the overall cost of the application to enhance the reliability by redundancy etc This firmware library provides a complete solution for sensorless detection of rotor position
31. of the speed feedback Such as disconnected wires Functions called None Input None Returns boolean TRUE a timeout occurred FALSE otherwise Ti 83 121 Library functions UM0492 84 121 HALL GetElectricalAngle Synopsis s16 HALL_GetElectricalAngle void Description This function exports the private variable containing the rotor electrical angle information In the present library this function is called by FOC algorithm since the rotor electrical angle is indispensable for performing Park transformation of stator currents Functions called None Input None Returns electrical angle s16 format See also Section 5 4 3 for detailed explanation about reference frame transformations HALL IncElectricalAngle Synopsis void HALL_IncElectricalAngle void Description As will be discussed later the software variable containing the rotor electrical angle information is synchronized with the feedback coming from the motor at each valid transition of the XOR of the three Hall sensor output In addition in order to increase the accuracy between two successive valid transitions the rotor electrical angle information is incremented each time the FOC algorithm is executed FOC_Module routine by accumulating the latest speed measurement dpp format The HALL_IncElectricalAngle function performs the accumulation of the speed and must consequently be called with the same sampling rate than the FOC algorithm Functions called
32. regulation of torque and flux stator current component Iq and Id during the alignment phase of the PMSM The function also updates the main state machine from Start to Run at the end of the alignment SVPWM_3ShuntGetPhaseCurrentValues or SVPWM_lIcsGetPhaseCurrentValues Clarke Park PID_Torque_Regulator PID_Flux_Regulator RevPark_Circle_Limitation SVPWM_IcsCalcDutyCycles or SVPWM_3ShuntCalcDutyCycles ENC_ResetEncoder ENC_Clear_Speed_Buffer None None Hall sensor feedback processing stm32 10x_hal11 module List of available functions The following is a list of available functions as listed in the stm32 10x_hal1 h header file HALL_HaliTimerInit on page 82 HALL_GetRotorFreq on page 82 HALL_GetSpeed on page 83 HALL_InitHallMeasure on page 83 HALL_IsTimedOut on page 83 HALL_GetElectricalAngle on page 84 HALL_IncElectricalAngle on page 84 HALL Init Electrical Angle on page 84 HALL_CirTimeOut on page 85 81 121 Library functions UM0492 82 121 HALL HallTimerInit Synopsis Description Functions called See also void HALL_HallTimerInit void The purpose of this function is to initialize the peripherals involved in Hall sensor feedback processing In particular GPIO input pins connected to the Hall sensors are initialized as floating inputs timer TIMx is configured in clear on capture mode and its XOR input function is enabled the prescaler is initialized with HALL_MAX_RATIO Fi
33. this function is to load the speed integral term with a default value UM0492 Library functions 5 8 2 PID regulator theoretical background The regulators implemented for Torque Flux and Speed are actually Proportional Integral Derivative PID regulators see note below regarding the derivative term PID regulator theory and tuning methods are subjects which have been extensively discussed in technical literature This section provides a basic reminder of the theory PID regulators are useful to maintain a level of torque flux or speed according to a desired target Figure 52 PID general equation torque f rotor position g torque and flux regulation for maximum flux f rotor position system efficiency E torque regulation for speed regulation torque f rotor speed E of the system Where Error sys Error of the system observed at time t T ErrOf sys Error of the system observed at time t T Tsampling T f X7 Kp x Error sys K x YErrorsys K x Errorgy Erroreys__ 1 0 x A Derivative term can be disabled Equation 1 corresponds to a classical PID implementation where K is the proportional coefficient Kiis the integral coefficient Kgis the differential coefficient Note As mentioned in Figure 52 the derivative term of the PID can be disabled independently through a compiler option see stm32 10x_MCconf h file for the torque flux or the speed regulation a PI can then be qui
34. used In order to reconstruct the currents flowing through a three phase load with the required accuracy using two ICSs it is necessary to properly synchronize A D conversions with the generated PWM signals This is why the two tasks are included in a single software module ky UM0492 Library functions 5 2 1 List of available functions and interrupt service routines The following is a list of available functions as listed in the stm32 10x_svpwm_ics h header file e SVPWM_kIcsInit on page 65 SVPWM_IcsCurrentReadingCalibration on page 66 e SVPWM_lIcsGetPhaseCurrentValues on page 66 SVPWM_lIcsCalcDutyCycles on page 67 SVPWM_Icslnit Synopsis Description Input Returns Note Functions called void SVPWM_Icslnit void The purpose of this function is to set up microcontroller peripherals for performing ICS reading and center aligned PWM generation The function initializes NVIC ADC GPIO and TIM1 peripherals In particular the ADC and TIM1 peripherals are configured to perform two pairs of simultaneous injected A D conversions every time the PWM registers are updated event called U event The first pair of conversions read the current values whereas the second pair acquires the bus voltage and the voltage at the temperature sensor Refer to Section 5 2 2 for further information on A D conversion triggering in ICS configuration None None It must be called at main level Standard library
35. variant Am compat ible yas Module4ocal symbols include all Cancel 13 121 Introduction to the sensorless FOC of PM motors UM0492 2 2 1 14 121 Introduction to the sensorless FOC of PM motors Introduction to the PM synchronous motor FOC drive This software library is designed to achieve the high dynamic performance in AC permanent magnet synchronous motor PMSM control offered by the well established field oriented control FOC strategy With this approach it can be stated that by controlling the two currents ig and igs which are mathematical transformations of the stator currents it is possible to offer electromagnetic torque Te regulation and to some extent flux weakening capability This resembles the favorable condition of a DC motor where those roles are held by the armature and field currents Therefore it is possible to say that FOC consists in controlling and orienting stator currents in phase and quadrature with the rotor flux this definition makes clear that a means of measuring stator currents and the rotor angle is needed Basic information on the algorithm structure and then on the library functions is represented in Figure 5 e the space vector PWM block CALC SVPWM implements an advanced modulation method that reduces current harmonics thus optimizing DC bus exploitation e the current reading block allows the system to measure stator currents correctly us
36. when the motor is running by using the LCD interface and evaluating the results If necessary the default value can be overwritten with a more suitable one e define PLL KI GAIN The default formula provides an a priori determination of the loop filter gain of the PLL Nonetheless this signed 16 bit value can be tuned when the motor is running by using the LCD interface and evaluating the results If necessary the default value can be overwritten with a more suitable one define F1 define F2 These coefficients signed 16 bit values help amplify the observer equations so that motor winding resistance and inductance can give a valuable contribution Maximum value is 215 32768 Depending on several system parameters motor parameters state observer parameters sampling frequency current and voltage conversion parameters compilation errors Pe068 or and Pe069 may occur in the MC_State_Observer_Interface c source file In that case the user should follow the procedure below to solve the problem Jump to the code line where the error was reported and see which Cx factor x 1 5 caused the issue Go to the definition of that Cx factor in MC_State_Observer_param h and assess whether F1 or F2 is involved Decrease F1 or F2 accordingly always choosing positive powers of two Startup parameters Rotor position reconstruction is based on the observation of the back emfs generated when the rotor is runni
37. with Hall sensors HALL_GetElectricalAngle if working in sensorless mode STO_Get_Electrical_Angle if working with ICS SVPWM_lIcsGetPhaseCurrentValues SVPWM_IcsCalcDutyCycles if working with three shunt SVPWM_3ShuntGetPhaseCurrentValues SVPWM_3ShuntCalcDutyCycles 69 121 Library functions UM0492 5 4 70 121 FOC_CalcFluxTorqueRef Synopsis Description Input Returns Functions called void FOC_CalcFluxTorqueRef void This function provides current components iqs and ids to be used as reference values by the FOC_Model1 function in Speed control mode see Torque and flux controller block in Figure 6 Speed setpoint and actual rotor speed are compared in a PID control loop whose output is iqs Speed setpoint is used to run through two lookup tables described in MC_PMSM_motor_param h Section 4 6 3 defined by taking into account the field weakening strategy explained in Section 2 3 in order to get the optimal flux reference ids and the saturation value of the torque current component iqs max that allow to reach the desired speed obviously for each rotor speed in the field weakening region there is a maximum torque achievable None None PID_Speed_Regulator if working with encoder ENC_Get_Mechanical_Speed if working with Hall sensors HALL_GetSpeed if working in sensorless mode STO_Get_Speed_Hz Reference frame transformations MC_Clarke P
38. 00 01000 Target 09300 Id Measured 00000 Move tlChange Moreover to achieve speed regulation in speed control mode a PI D is also implemented inside the torque and flux controller block The tuning of its related gains can be done in real time by means of the dedicated LCD menu Figure 19 LCD screen for setting the P term of the speed PID STM32 Motor Control PMSM FOC ver 1 90 Speed P I 01000 00700 Target 01500 rpm Measured oo0000 rpm Move tlChange Like for the previous menus either of the P or D when present coefficients can be selected highlighted in red by moving the joystick to the right left The desired values can then be changed incremented or decremented by pressing the joystick up down ky UM0492 Running the demo program 3 4 3 5 Observer and PLL gain tuning In the default configuration of the firmware library the tuning of the sensorless algorithm is disabled Nevertheless when the OBSERVER_GAIN_TUNING definition is not commented in the stm32F10x_MCconf h configuration header file a dedicated menu is shown on the LCD Figure 20 LCD screen for setting the P term of the flux PID STM32 Motor Control PMSM FOC ver 1 90 Observer Gains K 1 K2 01200 00852 PLL Gains P l 00237 00046 Move tlChange When the menu shown in Figure 20 is displayed the joystick can be moved to the right left to navigate between the di
39. 0492 Getting started with tools 1 4 2 Extended linker file setting As mentioned in the previous section in the provided IAR workspace the internal Flash extended linker file is set by default Inkarm_flash xcl To modify the linker file to be used for example Inkarm_ram xcl Open the IAR workspace by double clicking on the STM32_FOC_PMSM EWARM STM32_FOC_PMSM eww file 1 Go to the Project menu select Options The dialog box shown in Figure 4 is displayed In the Override default field type the name of the linker file you want to use and then segments on RAM memory 2 menu 3 click OK Selecting the Figure 4 Extended linker file setting then Linker and select the Config sub Inkarm_ram xcl file makes the IAR XLINK linker place the memory Category General Options C C Compiler Assembler Custom Build Build Actions Debugger Simulator Angel IAR ROM monitor J ink J Trace LMI FTDI Macraigor RDI Third Party Driver Options for node aSIM 32 FOG PMSM Factory Settings Output Extra Output define Diagnostics List Config Proce 4 gt r Output file Overide default STM32_FOC_PMSM d79 Secondary output file None for the selected format Debug information for C SPY IV With runtime control modules IV With 1 0 emulation modules Buffered terminal output IV Allow C SPY specific extra output file C Other Output format Format
40. 16 bit value adjustable from 0 to 32767 define PID_TORQUE_KD_DEFAULT The derivative constant of the torque loop regulation signed 16 bit value adjustable from 0 to 32767 define PID_FLUX_REFERENCE Flux lg reference its default value is 0 define PID_FLUX_KP_DEFAULT The proportional constant of the flux loop regulation signed 16 bit value adjustable from 0 to 32767 define PID_FLUX_KI_DEFAULT The integral constant of the flux loop regulation signed 16 bit value adjustable from 0 to 32767 define PID_FLUX_KD_DEFAULT The derivative constant of the flux loop regulation signed 16 bit value adjustable from 0 to 32767 define SQUARE _WAVE_PERIOD The period in ms of the square wave reference torque generated when FLUX_TORQUE_PIDs_TUNING is uncommented in stm32 10x_MCconf h 43 121 Getting started with the library UM0492 4 3 44 121 Linear variation of PID constants according to mechanical speed Refer to Section 5 8 4 Adjusting speed regulation loop Ki Kp and Kd vs motor frequency on page 98 Incremental encoder parameters MC_encoder_param h The MC_encoder_parameter h header file is to be filled if position speed sensing is performed by means of a quadrature square wave relative rotary encoder ENCODER or VIEW_ENCO
41. 2F10xFWLib Vx y The STM32 FOC Firmware Libraries v1 0 folder contains the firmware libraries for running 3 phase PMSM and AC sensored induction motors and 3 phase PMSM sensorless induction motors The STM32F10xFWLib Vx y folder contains the standard library for the STM32F103xx The inc sre and lib folders contain the header files the source files and the library files of the motor control library Finally EWARM folder contains the configuration files for the IAR toolchain Starting the IAR toolchain When you have installed the toolchain you can open the workspace directly from the dedicated folder by double clicking on the STM32_FOC_PMSM eww file The file location is STM32_FOC_PMSM EWARM STM32_FOC_PMSM eww 11 121 Getting started with tools UM0492 1 4 1 4 1 12 121 Customizing the workspace for your STM32F103xx derivative The PMSM FOC software library was written for the STM32F103VB6 However it works equally successfully with all the products in the STM32F103xx performance line family Using a different STM32F103xx sales type may require some modifications to the library according to the available features some of the I O ports are not present on low pin count packages Refer to the MCU datasheet for further details Also depending on the memory size the workspace may have to be configured to fit your STM32F103xx MCU derivative Inkarm_xxx xc1 file internal Flash or RAM based projec
42. 3ShuntCurrentReadingCalibration UM0492 Library functions SVPWM_3ShuntCurrentReadingCalibration Synopsis Description Input Returns Note Functions called void SVPWM_3ShuntCurrentReadingCalibration void The purpose of this function is to store the three analog voltages corresponding to zero current values for compensating the offset introduced by the amplification network None None This function reads the analog voltage on the ADC channels used for current reading For this reason it must be called before the PWM outputs are enabled so that the current flowing through the inverter is zero Those values are then stored into Phase_x_Offset variables Standard library ADC_ITConfig ADC_ExternalTrigInjectedConvConfig ADC_ExternalTrigInjectedConvCmd ADC_InjectedSequencerLengthContfig ADC_InjectedChannelConfig ADC_ClearFlag ADC_SoftwareStartInjectedConvCmd ADC_GetFlagStatus ADC_GetlnjectedConversionValue ADC_SoftwareStartInjectedConvCmd Motor control library SVPWM_InjectedConvConfig SVPWM_3ShuntGetPhaseCurrentValues Synopsis Description Input Returns Note Functions called Curr_Components SVPWM_3ShuntGetPhaseCurrentValues void This function computes current values of Phase A and Phase B in q1 15 format starting from values acquired from the A D Converter peripheral None Curr_Components type variable In order to have a q1 15 format for the current values the digital value
43. 6 bits None Link between Electrical Mechanical frequency RPM Electrical frequency number of pair poles x mechanical frequency RPM speed 60 x Mechanical frequency RPM revolutions per minute Example electrical frequency 100 Hz motor with 8 pair poles 100HZz electrical lt gt 100 8 12 5Hz mechanical lt gt 12 5 x 60 750 RPM 79 121 Library functions UM0492 ENC_ResetEncoder Synopsis void ENC_resetEncoder void Description This function writes into the encoder timer register the value corresponding to the alignment angle set in MC_encoder_param h It is called at the end of any alignment phase Functions called None ENC_Clear Speed _ Buffer Synopsis void ENC_Clear_Speed_Buffer void Description This function resets the buffer used for speed averaging Functions called None ENC_Get_ Mechanical Speed Synopsis s16 ENC_Get_Mechanical_Speed void Description This function returns the rotor speed in Hz The value returned is given with 0 1Hz resolution which means that 1234 is equal to 123 4 Hz Input None Output Signed 16 bits Functions called None Note This routine returns the mechanical frequency of the rotor To find the electrical speed use the following conversion electrical frequency number of pole pairs mechanical frequency ENC_Calc_Average_Speed Synopsis void ENC_Calc_Average_Speed void Description This function must be called every SPEED_MEAS_TIMEBASE ms it computes th
44. 9 Figure 20 Figure 21 Figure 22 Figure 23 Figure 24 Figure 25 Figure 26 Figure 27 Figure 28 Figure 29 Figure 30 Figure 31 Figure 32 Figure 33 Figure 34 Figure 35 Figure 36 Figure 37 Figure 38 Figure 39 Figure 40 Figure 41 Figure 42 Figure 43 Figure 44 Figure 45 Figure 46 Figure 47 Figure 48 ky Firmware architecture 0 eee 1 JTAG interface for debugging and programming 20 ccc eee eee eee 10 Pil SIrUCture aiei ak a wins he Ree ew ee eae Sa A eee eee ee ee A 11 Extended linker file setting 0 6 0 eect tenes 13 FOC algorithm structure torque control 2 2 2 0 eee 15 Speed Control loop eevee ee es a ee doh eee ead a ee ee Oe 15 Different PM motor constructions 000 cee tee ee 16 Assumed PMSM reference frame convention 0000 e cece ne een eee 18 General sensorless algorithm block diagram 000 eee eee 19 PMSM back emfs detected by the sensorless state observer algorithm 20 Flux weakening operation trajectory 0 0 ee tees 21 LCD screen for Torque control settings 0 00 ec e eee teeee 22 LCD screen for Target Iq settingS 0 0 0 00 cece eee 22 LCD screen for Target Id settingS 0 0 0 00 e eee 23 Speed control main settings 0600 c cette eee 24 LCD screen for setting Target speed 2 0 0 c eee 24 LCD screen for setting the P term of torque PID
45. AG interface for debugging and programming Using the JTAG interface of the MCU you can enter in circuit debugging session with most of toolchains Each toolchain can be provided with an interface connected between the PC and the target application Figure 2 JTAG interface for debugging and programming Application board MB525 The JTAG interface can also be used for in circuit programming of the MCU Other production programmers can be obtained from third parties UM0492 Getting started with tools 1 3 1 3 1 Note 1 3 2 1 3 3 Library source code Download The complete source files are available for free download on the ST website www stmcu com in the Technical Literature and Support Files section as a zip file It is highly recommended to check for the latest releases of the library before starting any new development and then to verify from time to time all release notes to be aware of any new features that might be of interest for your project Registration mechanisms are available on ST web sites to automatically obtain updates File structure The PMSM FOC software library contains the workspace for the IAR toolchain Once the files are unzipped the following library structure appears as shown in Figure 3 Figure 3 File structure O 132 FOC Firmware Libraries v1 0 E STM32_FOC_PMSM gt EWARM inc lib src 2 amp STM32_FOC_ACIM D 4 5 STM3
46. DER_FEEDBACK defined in stm32 10x_MCconf h define TIMER2_HANDLES_ENCODER To be uncommented if the two sensor output signals are wired to TIMER2 input pins default required when using STM3210B MCKIT e 8 define TIMER3 HANDLES ENCODER To be uncommented if the two sensor output signals are wired to TIMERS input pins e define TIMER4 HANDLES ENCODER To be uncommented if the two sensor output signals are wired to TIMER4 input pins e define ENCODER_PPR Defines the number of pulses generated by a single channel for one shaft revolution actual resolution will be 4x e define MINIMUM_MECHANICAL_SPEED_RPM Defines in rpm the minimum speed below which the speed measurement is either not realistic or not safe in the application an error counter is increased every time the measured speed is below the specified value In order to disable this check and control the motor down to zero speed the user can set this parameter to zero a warning message pointless comparison of unsigned integer with zero will be issued at compilation time e define MAXIMUM_MECHANICAL_SPEED_RPM Defines in rpm the maximum speed above which the speed measurement is either not realistic or not safe in the application an error counter is increased every time the measured speed is above the specified value e define MAXIMUM_ERROR_NUMBER Defines the number of consecutive errors on speed measurements to be detected before a fault event is generated check rate is specifi
47. DutyCycles on page 56 SVPWM_3ShuntAdvCurrentReading on page 56 53 121 Library functions UM0492 54 121 SVPWM_3Shuntinit Synopsis Description Input Returns Note Functions called void SVPWM_3ShuniInit void The purpose of this function is to set up microcontroller peripherals for performing 3 shunt resistor topology current reading and center aligned PWM generation The function initializes NVIC ADC GPIO TIM1 peripherals In particular the ADC and TIM1 peripherals are configured to perform two simultaneous A D conversions per PWM switching period Refer to Section 5 1 3 for further information None None It must be called at main level Standard library RCC_ADCCLKConfig RCC_AHBPeriphClockCmad RCC_APB2PeriphClockCmd GPIO_Structlnit GPIO_Init GPIO_PinLockConfig GPIO_PinRemapConfig TIM1_Delnit TIM1_TimeBaseStructinit TIM1_TimeBaselnit TIM1_OCStructinit TIM1_OC1Init TIM1_OC2Init TIM1_OC3lnit TIM1_OC4Init TIM1_OC1PreloadConfig TIM1_OC2PreloadConfig TIM1_OC3PreloadConfig TIM1_OC4PreloadConfig TIM1_BDTRConfig TIM1_SelectOutputTrigger TIM1_ClearlTPendingBit TIM1_ITConfig TIM1_Cmd TIM1_GenerateEvent TIM1_ClearFlag ADC_Delnit ADC_Cmd ADC_ Structlnit ADC_Init ADC_StartCalibration ADC_GetCalibrationStatus ADC_RegularChannelConfig ADC_InjectedSequencerLengthConfig ADC_InjectedChannelConfig NVIC_PriorityGroupConfig NVIC_Structlnit NVIC_Init Motor control library SVPWM_
48. EED_SENSORS To be uncommented to use the rotor position information inside the FOC algorithm rotor position information is provided by a state observer In this case the user should fill MC_State_Observer_param h and MC_PMSM_motor_param h in parallel as explained in Section 4 5 and Section 4 6 respectively define VIEW_HALL_FEEDBACK To be uncommented only in conjunction with NO_SPEED_SENSORS It is used to process the information that comes from three Hall sensors to be displayed through the DAC functionality and compared with the information coming from the sensorless rotor position reconstruction algorithm define VIEW_ENCODER_FEEDBACK To be uncommented only in conjunction with NO_SPEED_SENSORS It is used to process the information coming from an incremental encoder to be displayed through the DAC functionality and compared with the information coming from the sensorless rotor position reconstruction algorithm define BRAKE _RESISTOR To be uncommented to enable the software management of a resistive brake refer to Section 3 12 2 for more information about the hardware modifications to be applied to the MB459 board UM0492 Getting started with the library Caution In order to avoid any damage to the power st
49. Fowo Refresh_Rate 6 6 CPU Load 24 10 100 24 10 100 ky UM0492 Additional information A 2 Note Selecting the update repetition rate based on the PWM frequency for 3 shunt resistor configuration Beyond the well known trade off between acoustical noise and power dissipation consideration should be given to selecting the PWM switching frequency using the PMSM FOC software library As discussed in Section 5 1 4 on page 60 depending on the PWM switching frequency a limitation on the maximum applicable duty cycle could occur if using three shunt resistor configuration for current reading Table 2 PWM frequency vs maximum duty cycle relationship on page 64 summarizes the performance of the system when the software library is used in conjunction with STM3210B MCKIT hardware The MB459 board is an evaluation platform it is designed to support different motor driving topologies PMSM and AC induction and current reading strategies single and three shunt resistors Therefore the figures given in Table 2 on page 64 should be understood as a starting point and not as a best case Moreover in order to guarantee the proper working of the algorithm and be sure that the new computed duty cycles will be applied in the next PWM period it is always necessary to finish executing the FOC algorithm before the next PWM U event begins as shown in Figure 56 Figure 56 AD conversions for three shunt topology stator curr
50. HALL GetSpeed Synopsis s16 HALL_GetSpeed void Description This routine computes the rotor mechanical frequency in 0 1 Hz format starting from the array storing the latest period measurements and according to the following formula _ CKTIM 10 upp 37 POLE_PAIR_NUM captured value prescaler value Where CKTIM is the timer peripheral clock and POLE_PAIR_NUM is the number of pole pairs Be aware that returned value is zero if the prescaler is equal to the maximum or a timeout occurred and that excessive speed or high frequency glitches will result in a predefined value being returned HALL_MAX_SPEED Functions called GetAvrgHallPeriod or GetLastHallPeriod Input None Returns It returns the mechanical speed in 0 1Hz unit Format is s16 HALL InitHallMeasure Synopsis void HALL_InitHallMeasure void Description It clears software FIFO where latest speed information is pushed This function must be called before starting the motor to initialize the speed measurement process Functions called HALL_ClrCaptCounter TIM_SetCounter TIM _Cmd TIM_ITConfig Input None Returns None See also Section A 7 for more information about dpp unit HALL_IsTimedOut Synopsis bool HALL_IsTimedOut void Description This function simply returns the status of the boolean variable containing the speed measurement timeout status In the proposed firmware library this function is called in Run state by the main c to check for possible faults
51. IONALITY in stm32f10x_MCconfh Setting up the system when using a brake resistor Due to its physical construction a PM synchronous motor is able to transform kinetic energy into electrical energy just like a dynamo Under a limited number of conditions this property of PM synchronous motors has to be taken into consideration to avoid possible damage to the hardware system For instance a dangerous situation could arise when e The six inverter switches are opened and the motor is running at a speed higher than the nominal one In this case the amplitude of the line to line B emf generated is higher than the nominal bus voltage e The control tries to brake an energy transfer from the load to the board occurs Unless the used power system has regenerative capabilities in both of these situations the inverter bulk capacitor is charged Furthermore depending on the rotor speed with reference to the first situation or on the amount of energy transferred with reference to the second situation the voltage across the bulk capacitors could increase to a destructive level A strategy for somehow dissipating the generated electrical energy is thus necessary Different methods could be implemented to do so but one of them in particular the utilization of a brake resistor is supported by the library presented in this user manual If the motor is operated beyond the rated speed it is m
52. M1 Update TIM1 Update TIM1 Update ADC trigger ADC trigger ADC trigger ai14840 PMSM field oriented control MC_FOC_Drive module The MC_FOC_Drive module designed for permanent magnet synchronous motors provides at the core decoupled electromagnetic torque T regulation and to some extent flux weakening capability In addition it provides speed regulation by PID feedback control To operate it requires no adjustment with all of the selectable current or speed sensing configurations in accordance with the settings in the stm32 10x_MCconf h file isolated current sensing ICS three shunt resistor current sensing encoder position and speed sensing Hall sensor position and speed sensing sensorless position and speed detection The MC_FOC_Drive module handles several functions of other modules and has no direct access to the microcontroller peripheral registers List of available C functions FOC_Model on page 69 e FOC_CalcFluxTorqueRef on page 70 UM0492 Library functions FOC_Model Synopsis Description Input Returns Functions called void FOC_Model void The purpose of this function is to perform PMSM torque and flux regulation implementing the FOC algorithm Current commands ig and igg which under field oriented conditions can control machine torque and flux respectively are defined outside this function in Speed control mode they are provided by means of speed
53. None Input None Returns None HALL Init Electrical Angle Synopsis void HALL_Init_Electrical_Angle void Description Hall effect sensors are absolute and it is thus possible to reconstruct the rotor position by simply reading the set of their outputs This operating principle is utilized in this software function to initialize the software variable containing the present electrical angle before any motor startup The function acts by reading the state of H3 H2 and H1 signal task performed by private function ReadHallState and consequently initializing the software variable The maximum obtainable accuracy is 30 electrical degrees that is 30 POLE_PAIR_NUM mechanical degrees Input None Returns None UM0492 Library functions 5 6 2 HALL ClrTimeOut Synopsis HALL_ClrTimeOut Description This function sets to FALSE the boolean variable containing the timeout error flag indicating that information was lost or speed is decreasing sharply Functions called None Input None Returns None Speed measurement implementation Thanks to the STM32F103xx general purpose timer TIMx features it is very simple to interface the microcontroller with three Hall sensors In fact when the TI1S bit in the TIMx_CR2 register is set the three signals on the TIMx_CH1 TIMx_CH2 and TIMx_CH3 pins are XORed and the resulting signal is input to the logic performing TIMx input capture In this way the speed measurement is converte
54. PMSM_motor_param h see Section 4 6 3 Re build the project and download it on the STM32F103xx microcontroller Please be aware that this procedure should also be followed when the final target is a sensorless drive In fact receiving speed position information from any kind of sensor makes it possible to achieve a more precise customization and to tune the algorithm utilized for sensorless rotor position reconstruction Library configuration file stm32 10x_MCconf h The purpose of this file is to declare the compiler conditional compilation keys that are used throughout the entire library compilation process to Select which current measurement technique is actually in use the choice is between three shunt or ICS sensors according to availability Select which speed position sensor is actually used here the choice is between Hall sensors quadrature incremental encoder or no speed sensors depending on the requirements e Enable or disable the derivative action in the speed controller or in the current controllers in accordance with expected performance and code size e Enable or disable the brake resistor management depending on requirements on brake performance and on maximum speed e Enable or disable a virtual 2 channel DAC for real time tracing of the most important software variables For the best debug support user should keep this feature enabled e Enable or disable the execution of a specific software dedicated to t
55. ST_I_STARTUP In digits defines the initial amplitude of the three phase current system according to the formula below digit e define FINAL_I_STARTUP In digits defines the final amplitude of the three phase current system according to the above formula This amplitude should be chosen to generate an electromagnetic torque T that matches the estimated load applied define I_STARTUP_DURATION In milliseconds defines the time allowed to increase current amplitude linearly from initial to final amplitude e define MINIMUM_SPEED RPM In RPM defines the minimum speed at which the startup procedure may end if the rotor speed position detection is considered reliable to make way for normal operations Ti 49 121 Getting started with the library UM0492 4 5 3 4 6 4 6 1 50 121 Statistics parameters State observer output results are continuously monitored to statistically assess the reliability of the rotor speed position information supplied Since this technique is used as a fault detection system the parameters described in this section are used to set the desired safety level e define VARIANCE_THRESHOLD This parameter sets the threshold for the speed measurement variance The sensorless algorithm calculation is not considered reliable if the variance of the observed speed is greater than the desired percentage of the mean val
56. This evaluation platform is designed to support several motor driving topologies PMSM and AC induction and current reading strategies single and three shunt resistors Therefore the figures provided in should be understood as a starting point and not as a best case The RevPark_Circle_Limitation function performs the discussed stator voltage components saturation as illustrated in Figure 46 Figure 46 Circle limitation working principle S16_MAX r MMI S16_MAX ai14845 Va and V represent the saturated stator voltage components to be passed to the Reverse Park transformation function while Vg and Vq are the outputs of the PID current controllers From geometrical considerations it is possible to draw the following relationship _ vg MMI S16_MAX k v MMI S16_MAX oe q 3 Vv iy 77 121 Library functions UM0492 5 5 5 5 1 78 121 In order to speed up the computation of the above equations while keeping an adequate resolution the value MMI S16 MAX a is computed and stored in a look up table for different values of k Furthermore considering that MMI depends on the selected PWM frequency a number of look up tables are stored in MC_Clarke_Park c with MMI ranging from 91 to 100 Once you have selected the required PWM switching frequency you should uncomment the Max Modulation Index definition corresponding to the selected PWM frequency in the MC_Control_Param h definitio
57. Volt MCL_SetFault if using a brake resistor ADC_GetInjectedConversionValue MCL_Set_Brake_On MCL_Set_Brake_ Off if enabling the DAC functionality MCDAC_Update_Value MCDAC_Update_Output if using an encoder ENC_Start_Up and if using DAC ENC_Get_Electrical_Angle if using Hall sensors if using DAC HALL_IncElectricalAngle HALL_GetElectricalAngle if using the sensorless algorithm STO_Start_Up STO_Calc_Rotor_Angle MCL_Get_BusVolt and if using DAC STO_Get_Electrical_Angle STO_Get_wlalfa_est STO_Get_wlbeta_est STO_Get_wBemf_alfa_est STO_Get_wBemf_beta_est Section 5 1 and Section 5 2 for more details 109 121 Additional information UM0492 Appendix A Additional information A 1 Note 110 121 Adjusting CPU load related to FOC algorithm execution The advanced control timer TIM1 peripheral has the built in capability of updating PWM registers only after a given number of PWM semi periods This feature is handled by a programmable repetition counter It is particularly useful to adjust the CPU load related to FOC algorithm execution for a given PWM frequency refer to STM32F 103xx reference manual for more information on programmable repetition counter When using ICS the injected chain of conversions for current reading is directly triggered by a PWM register update event Moreover since the FOC algorithm is executed at the end of the injected chain of conversions in the related
58. _Speed on page 80 ENC_ErrorOnFeedback on page 81 ENC_Start_Up on page 81 UM0492 Library functions ENC_Init Synopsis Description Functions called See also void ENC_Init void The purpose of this function is to initialize the encoder timer The peripheral clock input pins and update interrupt are enabled The peripheral is configured in 4X mode which means that the counter is incremented decremented on the rising falling edges of both timer input 1 and 2 TIMx_CH1 and TIMx_CH2 pins RCC_APB1PeriphClockCmd RCC_APB2PeriphClockCmad GPIO_Structlnit GPIO_Init NVIC_Init TIM_Delnit TIM_TimeBaseStructInit TIM_TimeBaselnit TIM_EncoderlInterfaceConfig TIM_ICInit TIM_ClearFlag TIM_ITConfig TIM_Cmd STM32F103xx reference manual TIMx in encoder interface mode ENC_Get_ Electrical Angle Synopsis Description Input Output Functions called s16 ENC_Get_Electrical_Angle void This function returns the electrical angle in signed 16 bit format This routine returns 0 for 0 degrees 32768 S16_MIN for 180 degrees 32767 S16_MAX for 180 degrees None Signed 16 bits None ENC_Get_ Mechanical Angle Synopsis Description Input Output Functions called Note s16 ENC_Get_Mechanical_Angle void This function returns the mechanical angle in signed 16 bit format This routine returns 0 for 0 degrees 32768 S16_MIN for 180 degrees 32767 S16_MAX for 180 degrees None Signed 1
59. age it is mandatory to utilize the brake resistor feature for operation above the nominal speed define Id_Iq_DIFFERENTIAL TERM ENABLED To be uncommented when differential terms for torque and flux control loop regulation PID are enabled define SPEED_DIFFERENTIAL_TERM_ENABLED To be uncommented when differential term for speed control loop regulation PID is enabled define FLUX_TORQUE_PIDs_TUNING To be uncommented when a rotor position sensor is utilized It generates a software dedicated to torque and flux PID gain tuning Fill Mc_Control_Param h in parallel define OBSERVER_GAIN_TUNING If uncommented it enables the visualization on LCD of a menu dedicated to state observer and PLL gain tuning define DAC_FUNCTIONALITY To be uncommented to enable the DAC functionality Refer to Section 3 5 for more detailed information about this feature Once these settings have been done only the required blocks will be linked in the project this means that you do not need to exclude c files from the build Caution When using shunt resistors for current measurement ensure that the REP_RATE parameter in MC_Control_Param h is set properly see Section 4 2 and Section A 2 Selecting the update repetition rate based on the PWM frequency for 3 shunt resistor configuration on page 111 for details 4 2 Drive control parameters MC_Control_Param h The MC_Control_Param h header file gathers parameters related to Power device param
60. age 93 STO_Get_wBemf_beta_est on page 94 STO_Gains_Init on page 94 STO_Gains_Update on page 94 STO Init Synopsis void STO_Init void Description _ It initializes to proper values all the variables related to the state observer To be called once before every motor startup Input None Returns None Note In the demo program it is called during the INIT state ky 89 121 Library functions UM0492 90 121 STO StateObserverInterface Init Synopsis Description Input Returns Note void STO_StateObserverlInterface_Init void This function initializes the Sensorless algorithm according to motor parameters default state observer gain vector K1 K2 and PLL gains data are retrieved in the MC_State_Observer_param h MC_PMSM_motor_param h header files None None During runtime using the STO_Obs_Gains_Update function it is possible at any time to overwrite these initial settings modifying observer and PLL gains STO Obs Gains Update Synopsis Description Input Returns Note STO Calc _ Synopsis Description Input Returns Note void STO_Obs_Gains_Update void The purpose of this function is to modify the state observer and PLL gains previously set by STO_StateObserverlnterface_Init None None In the demo program by uncommenting OBSERVER_GAIN_TUNING in stm32f10xMCconf h it is possible through STO_Obs_Gains_Update to fine tune the sensorless algorithm
61. aken into account It is set by default to 2 if the timer overflows more than twice meaning that the Hall sensor period has been increased by a factor of at least two between two consecutive valid edges the number of overflows is not counted anymore This usually indicates that information has been lost Hall sensor timeout or that speed is decreasing very sharply The corresponding timeout delay depends on the selected timer prescaler which is variable the higher the prescaler low speed the longer the timeout period see also Section 5 6 define HALL SPEED _FIFO_SIZE This is the length of the software FIFO in which the latest speed measurements are stored This stack is necessary to compute rolling averages on several consecutive pieces of data 4 5 State observer parameters MC_State_Observer_param h The Mc_State_Observer_param his to be filled if the user wants to take advantage of or evaluate the implemented sensorless algorithm for rotor position speed detection In that case the MC_PMSM_motor_param h header file also has to be configured accordingly see Section 4 6 See also Section 3 11 which guides you progressively through the steps that make a sensorless system up and running The gathered parameters are related to the state observer startup measurement statistics and reliability 4 5 1 State observer parameters e define MAX CURRENT In Amperes defines the current value equivalent to an ADC conversion eq
62. andatory to use a regenerative power converter or a brake resistance to prevent bus overvoltage from damaging your board ky UM0492 Running the demo program 3 12 1 3 12 2 3 13 How to configure the FOC software library for brake resistor management To enable the management of the brake resistor on the STM32F103xx PMSM FOC firmware library simply uncomment the definition of BRAKE_RESISTOR in the stm32f10x_MCcont h header file The analog watchdog feature of the STM32F103xx allows to generate an interrupt whenever the bus voltage goes above the OVERVOLTAGE_THRESHOLD_V parameter specified in MC_Control_Param h and as a consequence the BRAKE_GPIO_PIN pin of the BRAKE_GPIO_PORT port is set to high level both definitions are in MC_MotorControl_Layer c default values are GPIOD and GPIO_Pin_13 for compatibility with STM3210B MCKIT From then on a hysteresis control of the bus voltage is performed Hysteresis can be entered by editing the BRAKE_HYSTERESIS parameter in stm32f10x_it c How to modify the MB459 board for brake resistor management In order to make the MB459 board suitable for the management of a brake resistor it is necessary to solder some additional components on its wrapping area Figure 28 gives an example of the circuit to be used for the hardware implementation of the brake Figure 28 Brake resistor circuit 15V 15V Vbus T T R3 1000h
63. ark module This module intended for AC machines induction synchronous and PMSM is designed to perform transformations of electric quantities between frames of reference that rotate at different speeds Based on the arbitrary reference frame theory the module provides three functions named after two pioneers of electric machine analysis E Clarke and R H Park These functions implement three variable changes that are required to carry out field oriented control FOC Clarke transforms stator currents to a stationary orthogonal reference frame named of frame see Figure 42 then from that arrangement Park transforms currents to a frame that rotates at an arbitrary speed which in PMSM field oriented control is synchronous with the rotor e Reverse Park transformation brings back stator voltages from a rotating frame q d to a stationary one The module also includes a function to calculate trigonometric functions sine and cosine and a function to correct the voltage vector command the so called Circle limitation ky UM0492 Library functions Figure 42 Clarke Park and reverse Park transformations Clarke Reverse Park gt ai14841 5 4 1 List of available C functions Clarke on page 72 Park on page 72 Rev_Park on page 72 Rev_Park_Circle_Limitation on page 73 Trig_Functions on page 73 iy 71 121 Library functions UM0492 72 121 Clarke Syn
64. brary functions MCL _ Chk _OverTemp Synopsis Description Input Returns bool MCL_Chk_OverTemp void This function performs the averaging of the latest temperature acquired value by means of the following formula XAV K XAV K 1 T_AV_ARRAY_SIZE 1 X K T_AV_ARRAY_SIZE where Xay K is the average at step K and X k the latest measurement at step K Once the average has been performed the function checks whether the acquired temperature is within the admitted range or not The intervention threshold and hysteresis values can be adjusted in MC_Control_Param h only for MB459 board None Returns TRUE if the software averaged voltage on the thermal resistor connected to ADC channel ADC_IN10 has reached the threshold level or if it has not yet returned to the threshold level minus the hysteresis value after an overheat detection Returns FALSE otherwise MCL Chk BusVolt Synopsis BusV_t MCL_Chk_BusVolt void Description This function checks for over and under voltage faults on inverter DC bus The intervention thresholds can be defined in MC_Control_Param h only for MB459 board Functions called None Input None Returns It returns a BusV_t type variable reporting the fault value MCL _Compute_BusVolt Synopsis Description Input Returns u16 MCL_Compute_BusVolt void This function computes the DC bus voltage in volt units In the proposed firmware library this function is utilized f
65. cal frequency rpm above which speed feedback is not realistic in the application used to discriminate glitches for instance define HALL MAX SPEED This parameter is the value returned by the HALL_Get Speed function 0 1 Hz unit if measured speed is greater than HALL_MAX_SPEED_FDBK_RPM The default value is 500 Hz but it can be 0 or FFFF depending on how this value is managed by the upper layer software define HALL _MAX PSEUDO_SPEED This parameter is the value returned by the HALL_GetRotorFreg function if measured speed is greater than MAX_SPEED_FDBK The unit is ddp See Appendix A 7 for more details about how to convert Hertz into pseudofrequency define HALL _MIN_SPEED_FDBK_RPM Defines the rotor mechanical frequency below which speed feedback is not realistic in the application define HALL _MAX RATIO It defines the lowest speed that can be measured when counter OxFFFF It also prevents the clock prescaler from decreasing excessively when the motor is stopped This prescaler is automatically adjusted during each capture interrupt to optimize the timing resolution UM0492 Getting started with the library define HALL_MAX_OVERFLOWS This is the maximum number of consecutive timer overflows t
66. ch off see Figure 39 Figure 39 ADutyA lt DT Ty Ts 2 and ADuty p gt DT Tpt Ts ADuty 2 Low side phase A e High side phase A a eee Low side phase B High side phase B PALA 1 rpt Current feedback phase B i f 1 e Sampling start Current feedback phase C bt re i Sampling end Space available for sampling P lt rad ai14838b Case 4 ADuty y lt DT Ty Ts 2 and ADuty p lt DT TRt Ts In this case the duty cycle applied to Phase A is so short that no current sampling can be performed between the two low side commutations Furthermore if the difference in duty cycles between Phases B and A is not long enough to allow the A D conversions to be performed between Phase B low side switch on and Phase A high side switch off it is impossible to sample the currents See Figure 40 To avoid this condition it is necessary to reduce the maximum modulation index or decrease the PWM frequency Figure 40 ADuty lt DT Ty Ts 2 and ADuty p lt DT TR Ts Piensa s ADuty 2 Low side phase A a High side phase A iB Low side phase B B High side phase B _ Tk iy 1 e i i Current feedback phase B E Current feedback phase C EOE ae No space available for sampling m i Fai lt T ai14839b ky 63 121 Library functions UM0492 5 2 64 121 The following parameters have been set as default in the firmware They are related to the MB459 board e DT 0 8 us
67. ckly implemented whenever the system doesn t require a PID control algorithm 5 8 3 Regulator sampling time setting The sampling time needs to be modified to adjust the regulation bandwidth As an accumulative term the integral term is used in the algorithm increasing the loop time decreases its effects accumulation is slower and the integral action on the output is delayed Inversely decreasing the loop time increases its effects accumulation is faster and the integral action on the output is increased This is why this parameter has to be adjusted prior to setting up any coefficient of the PID regulator In order to keep the CPU load as low as possible and as shown in equation 1 in Figure 52 the sampling time is directly part of the integral coefficient thus avoiding an extra multiplication Figure 53 describes the link between the time domain and the discrete system 97 121 Library functions UM0492 Note 5 8 4 98 121 Figure 53 Time domain to discrete PID equations Time domain f t Kp x ErroOfsys t K x J jE tOrgyg t dt Kyx Errorsys t Discrete f Xy K x Errorsys k x DYErrorsys Kgx Errors Errorsys domain 0 sampling done at Fs 1 Ts frequency pe kx To K In theory the higher the sampling rate the better the regulation In practice you must keep in mind that e The related CPU load will grow accordingly e For speed regulation there is absolutely no nee
68. d into the period measurement of a square wave having a frequency three times higher than the real electrical frequency The only exception is that the rolling direction which is not extractable from the XORed signal is on the contrary performed by directly accessing the three Hall sensor output Rolling direction identification As shown in Figure 47 it is possible to associate any of Hall sensor output combinations with a state whose number is obtainable by considering H3 H2 H1 as a three digit binary number H3 is the most significant bit Figure 47 Hall sensors output state correspondence 3 sensors 120 3 sensors 60 State5 gt State1 gt State 3 gt State1 gt State3 gt State 7 gt State2 gt State6 gt State 4 gt State6 gt State4 gt State 0 gt ai14846 Consequently it is possible to reconstruct the rolling direction of the rotor by comparing the present state with the previous one and considering that in presence of a positive speed the sequence must be the one illustrated in Figure 47 85 121 Library functions UM0492 86 121 Period measurement Although the principle for measuring a period with a timer is quite simple it is important to keep the best resolution in particular for signals such as the one under consideration that can vary with a ratio that can easily reach 1 1000 In order to always have the best resolution the timer clock pr
69. d to have a sampling time lower than the refresh rate of the speed information fed back by the external sensors this becomes especially true when Hall sensors are used while driving the motor at low speed As discussed in Section 4 2 on page 41 the speed regulation loop sampling time can be customized by editing the PID_SPEED_SAMPLING_TIME parameter in the MC_Control_Param h header file The flux and torque PID regulator sampling rates are given by the relationship 2 PWM FREQ Fl PID li ux and torque sampling rate REP RATE 1 REP_RATE must be an odd number if currents are measured by shunt resistors see also Section A 2 on page 111 its value is 8 bit long Adjusting speed regulation loop Ki Kp and Kd vs motor frequency Depending on the motor frequency it might be necessary to use different values of Kp Ki and Kd These values have to be input in the code to feed the regulation loop algorithm A function performing linear interpolation between four set points PID_Speed_Coefficient_update is provided as an example in the software library see MC_PID_regulators c and can be used in most cases as long as the coefficient values can be linearized If that is not possible a function with a larger number of set points or a look up table may be necessary To enter the four set points once the data are collected edit the MC_Control_param h file and fill in the field dedicated to the Ki Kp a
70. document supersedes and replaces all information previously supplied The ST logo is a registered trademark of STMicroelectronics All other names are the property of their respective owners 2008 STMicroelectronics All rights reserved STMicroelectronics group of companies Australia Belgium Brazil Canada China Czech Republic Finland France Germany Hong Kong India Israel Italy Japan Malaysia Malta Morocco Singapore Spain Sweden Switzerland United Kingdom United States of America www st com 121 121 z
71. e is out of range or an error code is returned Note Indicates the limits of the function or specific requirements that must be taken into account before implementation Caution Indicates important points that must be taken into account to prevent hardware failures Functions called Lists called functions Useful to prevent conflicts due to the simultaneous use of resources Code example Indicates the proper way to use the function and if there are certain prerequisites interrupt enabled etc Some of these sections may not be included if not applicable for example no parameters or obvious use Current reading in three shunt resistor topology and space vector PWM generation stm32 10x_svpwm_3shunt module Two important tasks are performed in the stm32 10x_svpwm_3shunt module Space vector pulse width modulation SVPWM Current reading in three shunt resistor topology In order to reconstruct the currents flowing through a three phase load with the required accuracy using three shunt resistors it is necessary to properly synchronize A D conversions with the generated PWM signals This is why the two tasks are included in a single software module List of available functions The following is a list of available functions as listed in the stm32f10x_svpwm_3shunt h header file e SVPWM_3Shuntinit on page 54 SVPWM_3ShuntCurrentReadingCalibration on page 55 SVPWM_3ShuntGetPhaseCurrentValues on page 55 SVPWM_3ShuntCalc
72. e PID SPEED _SAMPLING_TIME parameter It undertakes two actions it averages the buffered values of observed speed storing the result ina module private variable e it calculates the population variance of that speed buffer if the variance is higher than the threshold settled by VARIANCE_THRESHOLD see Section 4 5 3 then speed estimation is declared not reliable and a module private flag is raised None None None STO InitSpeedBuffer Synopsis Description Input Returns Note void STO_InitSpeedBuffer void This function initializes the buffer used by STO_Calc_Speed to store observed rotor speed To be called once before every motor startup None None In the demo program it is called during the WAIT and FAULT states STO Get_Electrical_Angle Synopsis Description Input Returns Note s16 STO_Get_Electrical_Angle void It returns the rotor electrical angle at step k 1 as STO_Calc_Rotor_Angle calculated and stored in a module private variable at time k None Observed rotor electrical angle s16 format None STO Get_Mechanical_Angle Synopsis Description Input Returns Note s16 STO_Get_Mechanical_Angle void It returns the rotor mechanical angle at step k 1 None Observed rotor mechanical angle s16 format This function relies on STO_Get_Electrical_Angle 91 121 Library functions UM0492 92 121 STO _Get_Speed Synopsis Description Input Returns No
73. e devices capable of sensing the polarity of the rotor s magnetic field they provide a logic output which is 0 or 1 depending on the magnetic pole they face and thus on the rotor position Typically in a three phase PM motor three Hall effect sensors are used to feed back the rotor position information They are usually mechanically displaced by either 120 or 60 and the presented firmware library was designed to support both possibilities To set up the PMSM FOC software library for use with three Hall sensors simply modify the stm32f10x_MCconf h and MC_hall_param h header files according to the indications given in Section 4 1 and Section 4 4 respectively As shown in Figure 26 the typical waveforms can be visualized at the sensor outputs in case of 60 and 120 displaced Hall sensors More particularly Figure 26 refers to an electrical period i e one mechanical revolution in case of one pole pair motor Figure 26 60 and 120 displaced Hall sensor output waveforms 3 sensors 120 3 sensors 60 ai14826 33 121 Running the demo program UM0492 Since the rotor position information they provide is absolute there is no need for any initial rotor prepositioning Particular attention must be paid however when connecting the sensors to the proper microcontroller inputs In fact as stated in Section 3 9 this software library assumes that the positive rolling direction is the rolling direction of a
74. e latest speed measurement if it is out of the range specified in MC_encoder_param h then the error counter is incremented and the speed is saturated Furthermore if the error counter is higher than MAXTMUM_ERROR_NUMBER the boolean variable storing the error status is set Finally the new average value is computed based on the latest SPEED_BUFFER_SIZE speed measurement The user can disregard the warning message pointless comparison of unsigned integer with zero that is issued by the compiler if MINIMUM _ MECHANICAL _SPEED_RPM has been set to zero Functions called ENC_Calc_Rot_Speed Input None Returns None 80 121 ky UM0492 Library functions 5 6 5 6 1 ENC_ErrorOnFeedback Synopsis Description Functions called Input Returns ENC_Start_Up Synopsis Description Functions called Input Returns bool ENC_ErrorOnFeedback void This function simply returns the status of the boolean variable containing the speed measurement error status which is updated every SPEED_MEAS_TIMEBASE ms by the ENC_Calc_Average_Speed function In the proposed firmware library this function is called in Run state by the main to check for possible faults of the speed feedback such as disconnected encoder wires None None boolean TRUE if an error occurred FALSE otherwise void ENC_Start_Up void The purpose of this function is to perform the
75. e phase load current has been updated and the FOC algorithm can then be executed in the A D end of injected conversion interrupt service routine JEOC ISR After execution of the FOC algorithm the value to be loaded into the OCR4 register is calculated to set the sampling point for the next PWM period and the A D converter is configured to sample the correct channels 59 121 Library functions UM0492 Note 60 121 Regular conversions are reserved for user purposes and must be configured manually See also firmware standard library user manual UM0427 Tuning delay parameters and sampling stator currents in three shunt resistor topology Figure 35 shows one of the three inverter legs with the related shunt resistor Figure 35 Inverter leg and shunt resistor position To indirectly measure the phase current I it is possible to read the voltage V providing that the current flows through the shunt resistor R It is possible to demonstrate that whatever the direction of current I it always flows through the resistor R if transistor T2 is switched on and T1 is switched off This implies that in order to properly reconstruct the current flowing through one of the inverter legs it is necessary to properly synchronize the conversion start with the generated PWM signals This also means that current reading cannot be performed on a phase where the duty cycle applied to the low side transistor is either null or very sh
76. ection 3 12 for more detailed information on how to set up your system when using brake resistor 5 11 Main interrupt service routines stm32 10x_it module The stm32 10x_it module can be used to describe all the exception subroutines that might occur within your application When an interrupt happens the software will automatically branch to the corresponding routine accordingly with the interrupt vector table With the exception of the ADC and TIM1 emergency stop break counter overflow and update if in shunt current reading mode interrupt requests all the routines are empty so that you can write your own code for exception handlers and peripheral interrupt requests ky 107 121 Library functions UM0492 5 11 1 108 121 List of non empty interrupt service routines As mentioned above only three interrupts are managed by motor control tasks TIM1_BRK_IRQHandler on page 108 e TIM1_UP_IRQHandler on page 108 ADC_IRQHandler on page 109 TIM1 BRK IRQHandler Synopsis Description Input Returns Functions called See also void TIM1_BRK_IRQHandler void The purpose of this function is to manage a break event on the dedicated BREAK pin In particular TIM1 outputs are disabled the main state machine is put into FAULT state None None MCL_SetFault TIM1_ClearPendingBit Advanced control timer TIM1 in STM32F103xx reference manual TIM1 UP_IRQHandler Synopsis Description Input Returns
77. ecution when REP_RATE 3 and PWM frequency gt 18 kHz 110 AD conversions for three shunt topology stator currents reading and FOC algorithm execution 0 ete 111 Block diagram of Pl controller 22 0 0 0 ee eae 113 Closed loop block diagram n snena ccc ete 114 Pole zero cancellation 2 0 2 2 0000 ccc eee 114 Block diagram of closed loop system after pole zero cancellation 115 KP 8000 and KI 2000 1 1 tte eae 117 KP 8000 and KI 1000 2 2 0 cece 118 UM0492 PMSM FOC software library V1 0 features CPU running at 72 MHz Supported speed feedbacks Sensorless 60 or 120 displaced Hall sensors Quadrature incremental encoder Current sampling method 2 isolated current sensors ICS 3 shunt resistors placed on the bottom of the three inverter lags DAC functionality for tracing the most important software variables Brake resistor management Speed control mode for speed regulation Torque control mode for torque regulation Field weakening 16 bit space vector PWM frequency can be easily adjusted Centered PWM pattern type 11 bit resolution at 17 6 kHz Rules for the a priori determination of all the parameters necessary for firmware customization CPU load below 25 in sensorless configuration 10 kHz FOC sampling rate Code size in sensorless configuration is about 13 Kbytes 3 shunt resistor current reading
78. ed by SPEED_MEAS_TIMEBASE in stm32 10x_Timebase c e define SPEED_BUFFER_SIZE Defines the buffer size utilized for averaging speed measurement Power of two is desirable for ease the computation Alignment settings Quadrature encoder is a relative position sensor Considering that absolute information is required for performing field oriented control it is necessary to somehow establish a 0 position This task is performed by means of an alignment phase and is carried out at first motor startup and after any fault event It basically consists in imposing a null reference torque lq and a reference flux Id with a linearly increasing magnitude and a constant orientation ky UM0492 Getting started with the library If properly configured at the end of this phase the rotor is locked in a well known position and the encoder timer counter is initialized accordingly The following parameters are used to customize the alignment phase depending on the motor inertia and load conditions T_ALIGNMENT in milliseconds defines the desired duration of the alignment phase ALIGNMENT_ANGLE specifies the vector orientation angle 8 in the below diagram Figure 30 Alignment angle b ai14833 e _ALIGNMENT in digits defines the final value of the reference Id magnitude With STM3210B MCKIT and with ALIGNMENT_ANGLE set to 90
79. ell placed on the a axis the green sinewave is current i C4 ky 19 121 Introduction to the sensorless FOC of PM motors UM0492 2 3 20 121 Figure 10 PMSM back emfs detected by the sensorless state observer algorithm More information on how to set parameters to make the firmware suit the user s motor could be found in Section 4 5 Introduction to flux weakening control The purpose of the flux weakening functionality is to expand the operating limits of a permanent magnet motor by reaching speeds higher than rated as many applications require under operating conditions where the load is lower than rated Here the rated speed is considered to be the highest speed at which the motor can still deliver maximum torque The magnetic flux can be weaken y acting on the direct axis current ig given a motor rated current lp such that lic ig if we choose to set id 0 then the maximum available quadrature current ig is reduced consequently in case of SM PMSM as shown in Section 2 1 2 the maximum deliverable electromagnetic torque is also reduced To optimize the efficiency of the operations theimplemented flux weakening strategy aims at driving the motor along a maximum torque per ampere current vector trajectory see 3 in Appendix A 8 References UM0492 Introduction to the sensorless FOC of PM motors Figure 11 Flux weakening operation trajectory voltage limit circles i bus voltage mo
80. en enabled in stm32_MCconf h this function initializes the GPIO pin driving the switch for resistive brake implementation BRAKE_GPIO_PORT BRAKE_GPIO_PIN are defined in MC_MotorControl_Layer c To be called after MCU reset Input None Returns None See also Section 3 12 for more detailed information on how to set up your system when using brake resistor UM0492 Library functions MCL Set _Brake_On Synopsis void MCL_Set_Brake_On void Description Declared and defined only if the brake resistor feature has been enabled in stm32_MCconf h it switches on the brake resistor by setting the BRAKE_GPIO_PIN pin of the BRAKE_GPIO_PORT port The function is called in ADC_IRQHandler every time an analog watchdog interrupt occurs Input None Returns None See also Section 3 12 for more detailed information on how to set up your system when using brake resistor MCL Set _Brake_ Off Synopsis void MCL_Set_Brake_Off void Description Declared and defined only if the brake resistor feature has been enabled in stm32_MCconf h this switch off the brake resistor by setting to zero the pin BRAKE_GPIO_PIN of port BRAKE_GPIO_PORT The function is called in ADC_IRQHandler if both the brake was turned on and the bus voltage went down the threshold specified by BRAKE_HYSTERESIS Note BRAKE_HYSTERESTS is defined in stm32f10x_it c and its default value is 15 16 the over voltage intervention threshold Input None Returns None See also S
81. ensorless FOC 0 00 e eee ee 51 4 6 3 Additional parameters for flux weakening operation 51 5 Library functions 6saaviobousesteese cd castcnbbwsed eee hedcnn 53 5 1 Current reading in three shunt resistor topology and space vector PWM generation stm32f10x_svpwm_3shunt module 53 5 1 1 List of available functions 000 00 e eee eee 53 5 1 2 Space vector PWM implementation 0 200 eee eee 57 5 1 3 Current sampling in three shunt topology and general purpose A D CONVEISIONS o eiur edari WE de Rae eee ee eee NG eS 58 5 1 4 Tuning delay parameters and sampling stator currents in three shunt resistor topology 60 5 2 Isolated current sensor reading and space vector PWM generation stm32f10x_svpwm_ics module 64 5 2 1 List of available functions and interrupt service routines 65 3 121 Contents UM0492 5 2 2 Current sampling in isolated current sensor topology and integrating general purpose A D conversions 000 e eee eee 67 5 3 PMSM field oriented control MC_FOC_Drive module 68 5 3 1 List of available C functions 0 000 cece e eee eee 68 5 4 Reference frame transformations MC_Clarke_Park module 70 5 4 1 List of available C functions 0 0 0 0 ccc eee eee 71 5 4 2 Detailed explanation about reference frame transformations 75 5 4 3 Circle limitation sic cacs eae sea OE Pad ada 76 5
82. ents reading and FOC algorithm execution New duty cycles ready before next U event OK New duty cycles not ready before next U event necessary to increase REP_RATE PWM Period l i ae U U7 REP_RATE 1 REP_RATE 1 LJ ADC triggering interval m FOC algorithm execution New duty cycle values ready ai14848 Considered that as seen in Section 5 1 4 the ADC conversions are triggered latest TN TS 2 after the TIM1 counter overflow and considered the time required for the A D converter to perform injected conversions it can been stated that the FOC algorithm is started about 5 us after the TIM1 counter overflow worst case Furthermore given that the execution time of the FOC algorithm is around 24 us in sensorless worst configuration to compute the new duty cycle values before the next update event it is necessary to guarantee a minimum duty cycle period of about 5 24 x 2 us that is a maximum achievable FOC execution rate of about 17 kHz Table 4 lists the minimum required repetition counter values as a function of the PWM frequency 111 121 Additional information UM0492 A 3 112 121 Table 4 System performance when using STM3210B MCKIT PWM frequency Max duty cycle Max FOC execution rate Up to 11 4 kHz 100 12 2 kHz 99 12 9 kHz 98 5 13 7 kHz 98 Equal to PWM frequency REP_RATE 1 14 4 kHz 98 15 2 kHz 97 16 kHz 96 5 16 7 kHz 96 5 PENE ee Equal to PWM
83. erence manual for more detailed information This case typically occurs when SVPWM with low lt 60 modulation index is generated see Figure 36 The modulation index is the applied phase voltage magnitude expressed as a percentage of the maximum applicable phase voltage the duty cycle ranges from 0 to 100 Figure 37 offers a reconstruction of the PWM signals applied to low side switches of Phase A and B in these conditions plus a view of the analog voltages measured on the STM32F103xx A D converter pins for both Phase B and C the time base is lower than the PWM period Figure 36 Low side switches gate signals low modulation indexes 100 90 80 70 60 50 40 30 0 30 60 Phase A Phase B Phase C Note that these current feedbacks are constant in the view in Figure 37 because it is assumed that commutations on Phase B and C have occurred out of the visualized time window Moreover it can be observed that in this case the two stator current sampling conversions can be performed synchronized with the counter overflow as shown in Figure 37 61 121 Library functions UM0492 Figure 37 Low side Phase A duty cycle gt DT Ty ADuty d Counter overflow e a Low side phase A pR i High side phase A r Low side phase B or High side phase B rat iT t Current feedback phase eE AEA a g SPREE e Sampling sta Current feedback phaseC ply os e Sam
84. escaler is constantly adjusted in the current implementation The basic principle is to speed up the timer if the captured values are too low for an example of short periods see Figure 48 and to slow it down when the timer overflows between two consecutive captures see example of large periods in Figure 49 Figure 48 Hall sensor timer interface prescaler decrease Capture Capture Transfer frompreload to active an next capture update evert Figure 49 Hall sensor timer interface prescaler increase Transfer from preload to active an next capture update event The prescaler modification is done in the capture interrupt taking advantage of the buffered registers the new prescaler value is taken into account only on the next capture event by the hardware without disturbing the measurement ky UM0492 Library functions Further details are provided in the flowchart shown in Figure 50 which summarizes the actions taken into the TIMx_IRQHandler Figure 50 TIMx_IRQHandler flowchart Capture event Read present state Extract spinning direction Synchronize angle Update OVF event OVF_Counter OVF_Counter gt HALL_MAX_OV ERFLOWS Time out TRUE Store capture prescaler direction Store capture prescaler direction Store capture presca
85. eters on page 41 Current regulation parameters on page 42 Power board protection thresholds on page 42 Speed loop sampling time on page 42 Speed P D controller init values on page 43 Quadrature current PID controller init values on page 43 Linear variation of PID constants according to mechanical speed on page 44 Power device parameters define PWM_FREQ Define here in Hz the switching frequency in parallel uncomment the maximum allowed modulation index definition MAX_MODULATION_XX_PER_CENT corresponding to the PWM frequency selection define DEADTIME_NS Define here in ns the dead time in order to avoid shoot through conditions 41 121 Getting started with the library UM0492 Note 42 121 Current regulation parameters e define REP_RATE Stator currents sampling frequency and consequently flux and torque PID regulators sampling rate are defined according to the following equation 2 PWM _FREQ Fl PID l 2 ux and torque sampling rate REP RATE 1 In fact because there is no reason for either executing the FOC algorithm without updating the stator currents values or for performing stator current conversions without running the FOC algorithm in the proposed implementation the stator current sampling frequency and the FOC algorithm execution rate coincide REP_RATE must be an odd number if currents are measured by shunt resistors see Section A 2 Selecti
86. etting up the system when using ICS sensors 2055 31 2 121 Ti UM0492 Contents 3 8 1 Selecting PHASE_A_ADC_CHANNEL and PHASE_B_ADC_CHANNEL 0 000 c eee eee 31 3 9 Setting up the system when using an encoder 005 32 3 10 Setting up the system when using Hall effect sensors 33 3 11 Progressive sensorless system development 0005 35 3 12 Setting up the system when using a brake resistor 36 3 12 1 How to configure the FOC software library for brake resistor management ce eee 37 3 12 2 How to modify the MB459 board for brake resistor management 37 3 13 Note on debugging tools 2 02 37 4 Getting started with the library 0 22 c eee eee eee 39 4 1 Library configuration file stm32 10x_MCconf h 39 4 2 Drive control parameters MC_Control_Param h 41 4 3 Incremental encoder parameters MC_encoder_param h 44 4 4 Hall sensor parameters MC_hall_prm h 200000000e 46 4 5 State observer parameters MC_State_Observer_param h 47 4 5 1 State observer parameters 0 00 47 4 5 2 Startup parameters 0 0 0 0 ene nas 48 4 5 3 Statistics parameters n wwe irae Pe Ahahe de Paes ae Ee 50 4 6 Permanent magnet synchronous motor parameters MC_PMSM_motor_param h 50 4 6 1 Basic motor parameters 00 0c eae 50 4 6 2 Motor parameters for s
87. fferent gains Pressing the joystick up down will increment decrement the gain highlighted in red color This menu is used to change both the observer and the PLL gains in real time This feature is particularly useful when used in conjunction with the DAC functionality and with a firmware configuration handling either Hall effect sensors or an encoder In this way it is possible to modify the observer and PLL gains by looking for example at both the observed and measured rotor electrical angle and by adjusting the gains so as to cancel any error between the two waveforms DAC functionality When enabled in the stm32F10x_MCconfh the DAC functionality is a powerful debug tool which allows the simultaneous tracing of up to two software variables selectable in real time using a dedicated menu Figure 21 LCD screen for setting the P term of the flux PID STM32 Motor Control PMSM FOC ver 1 90 Signal on PBO Observed el angle Signal on PB1 Measured el angle Move tlChange 27 121 Running the demo program UM0492 3 6 3 7 28 121 When the menu shown in Figure 21 is displayed the joystick can be moved to the right left to select the desired microcontroller pin To change the software variable in output move the joystick up down the list of the available variables depends on the selected firmware configuration For all other menus pressing the joystick or the Key button will cause the motor to start stop
88. g TIM1_ITConfig UM0492 Library functions 5 1 2 Space vector PWM implementation Figure 32 shows the stator voltage components Va and Vg while Figure 33 illustrates the corresponding PWM for each of the six space vector sectors Figure 32 V and Vz stator voltage components 0 5 0 3 0 1 Vbeta Valfa Figure 33 SVPWM phase voltages waveforms 100 A ana as 90 80 70 60 50 40 30 20 10 0 awe alin ai 0 60 120 180 240 300 360 Phase A Phase B Phase C Sector 1 Sector2 Sector 3 Sector 4 Sector5 Sector 6 57 121 Library functions UM0492 58 121 With the following definitions for U 3x Tx V Us Tx Vg and X Us Q amp I y 2508 and z Boe literature demonstrates that the space vector sector is identified by the conditions shown in Table 1 Table 1 Sector identification Y lt 0 Y gt 0 Z lt 0 Z gt 0 Z lt 0 Z gt 0 X lt 0 X gt 0 X lt 0 X gt 0 Sector V IV III VI II The duration of the positive pulse widths for the PWM applied on Phase A B and C are respectively computed by the following relationships _T X Z Sector I IV t 5 tg tatZ te tg X T Y Z Sector Il V ta 5 te tat Zs te ta Y Sector Ill VI t p LEAT tp te X te ta Y where T is the PWM period VE tA ge iB totA te any p Now considering that the PWM pattern is center aligned and
89. g through a three phase system follow the mathematical relationship l4 lo 1g 0 Therefore to reconstruct the currents flowing through a generic three phase load it is sufficient to sample only two out of the three currents while the third one can be computed by using the above relationship The flexibility of the STM32F103xx A D converter trigger makes it possible to synchronize the two A D conversions necessary for reconstructing the stator currents flowing through the motor with the PWM reload register updates The update rate can be adjusted using the repetition counter This is important because as shown in Figure 41 it is precisely during counter overflow and underflow that the average level of current is equal to the sampled current Refer to the STM32F103xx reference manual to learn more about A D conversion triggering and the repetition counter Note Regular conversions are reserved for the user and must be configured manually See also firmware standard library user manual UM0427 iy 67 121 Library functions UM0492 5 3 5 3 1 68 121 Figure 41 Stator currents sampling in ICS configuration REP_RATE 1 PWM Counter Compare A a i a aa a es A a ee a NS S Compare B E es EE E Aan AT E E Low side A Low side B A E OEE AAN PAE AE Qpar Phase current i et i i 1 i TI
90. h from the Fault to the Idle state There are six different fault sources when using the firmware library in conjunction with the STM3210B MCKIT Overcurrent A low level was detected on the PWM peripheral dedicated pin BKIN If the STM3210B MCKIT is being used this means that either the hardware overtemperature protection or the hardware overcurrent protection has been triggered Refer to the STM3210B MCKIT user manual for details Overheating An overtemperature was detected on the dedicated analog channel The intervention threshold NTC_THRESHOLD_C and the related hysteresis NTC_HYSTERESIS_C are specified in the MC_Control_Param h header file Refer to the STM3210B MCKIT user manual for details Bus overvoltage Available only if the BRAKE RESISTOR definition is commented default in stm32f10x_MCconf h configuration header file It means that an overvoltage was detected on the dedicated analog channel The intervention threshold OVERVOLTAGE_THRESHOLD_V is specified in the MC_Control_Param h header file Refer to STM3210B MCKIT user manual for details If the BRAKE_RESISTOR definition is not commented in stm32f10x_MCcont h it is assumed that a resistor with a high power dissipation capability was connected in parallel to the bus capacitors through a switch In this case the overvoltage does not generate a FAULT event because the resistor is supposedly able to dissipate the excess of voltage across the bus ca
91. he MC_Control_Param h file see Section 4 2 on page 41 PID Flux Regulator Synopsis Description Input Output Functions called Note See also s16 PID_Flux_regulator PID_FluxTYPEDEF PID_Flux s16 qld_input The purpose of this function is to compute the proportional integral and derivative terms if enabled see Id_Iq_DIFFERENTIAL_TERM_ENABLED in Section 4 1 on page 39 for flux Ips regulation PID_FluxTYPDEF see MC_type h for structure declaration signed 16 bits Signed 16 bits None Default values for the PID flux regulation are declared and can be modified in the MC_Control_Param h file see Section 4 2 on page 41 Figure 58 on page 114 shows the PID block diagram PID Torque Regulator Synopsis Description Input Output Functions called Note See also s16 PID_Torque_regulator PID_TorqueT YPEDEF PID_Torque s16 qlq_input The purpose of this function is to compute the proportional integral and derivative terms if enabled see Id_lq_DIFFERENTIAL_TERM_ENABLED in Section 4 1 on page 39 for the torque los regulation PID_TorqueTYPDEF see MC_type h for structure declaration signed 16 bits signed 16 bits None Default values for the PID torque regulation are declared and can be modified in the MC_Control_Param h file see Section 4 2 on page 41 Figure 58 on page 114 shows the PID block diagram 95 121 Library functions UM0492 96 121 PID Speed_Regulator Synopsis
92. he tuning of current PIDs e Enable or disable the tuning of the State Observer and PLL gains e Incase of no speed sensor it is still possible to acquire signals coming from Hall sensors or encoder and evaluate whether the sensorless algorithm is working properly using the DAC functionality Those signal will not be used in the FOC algorithm the choice here is between Hall sensors quadrature incremental encoder or none 39 121 Getting started with the library UM0492 40 121 If this header file is not edited appropriately no choice or undefined choice you will receive an error message when building the project Note that you will not receive an error message if the configuration described in this header file does not match the hardware that is actually in use or in case of wrong wiring More specifically define ICS_SENSORS To be uncommented when current sampling is done using isolated current sensors define THREE_SHUNT To be uncommented when current sampling is performed via three shunt resistors default define ENCODER To be uncommented when an incremental encoder is connected to the starter kit for position sensing in parallel fill out MC_encoder_param h as explained in Section 4 3 define HALL SENSORS To be uncommented when three Hall sensors 60 or 120 displaced are in use to detect rotor speed default in parallel fill out MC_hall_prm h as explained in Section 4 4 define NO_SP
93. he user can for instance look at the real and measured Iq current using DAC functionality and slightly change the torque PI D gains with respect to the default values in order to have a quick response to a step of target Iq without overshoots The same PI D gains could then also be used in the flux loop Figure 61 and Figure 62 show two oscilloscope acquisition In both acquisitions the C1 channel is the PBO pin output and the C2 channel is the PB1 pin output The DAC functionality was used to output two internal variables in this case PBO is the reference lq and PB1 is the measured lq The PBO and PB1 signals were analog and digital filtered It is possible to see in Figure 67 that the measured Iq has an overshoot with respect to the reference lq In this case the PI parameter was set to KP 8000 and KI 2000 To reduce this overshoot it is required to decrease KI while keeping KP constant This is how the condition of Figure 62 with KP 8000 and KI 1000 is obtained It is possible to see that in Figure 62 the response of the current control is slower than in Figure 62 So it is possible to increase the speed of the system by increasing KP and KI and keeping the ratio constant UM0492 Additional information Figure 61 KP 8000 and KI 2000 eoo cz ry Measure P1irms C4 P2 top C4 P3 rms C4 P4freq C2 PS P amp value Tany 1 mY status 7 A 200 mV idiv 200 m idiv 2 00
94. ied by ky UM0492 Running the demo program moving the joystick up down Once the motor Start command has been issued the Target Id can be changed in runtime while the measured Id current component is shown in the Measured field Figure 14 LCD screen for Target Id settings STMZ2 Motor Control PMSM FOC ver 1 0 Sensorless Demo Torque control mode Target Measured Iq 0323000 00000 Id 09334 00000 Speed 00000 Move tlChange The motor is stopped main state machine moves from Run to Stop state by pressing either the KEY button or the joystick Different motor ramp up strategies are used in torque control mode depending on the kind of configuration utilized for the speed position feedback ENCODER Or VIEW_ENCODER_FEEDBACK uncommented in the configuration file stm32f10x_MCcon fh In this case a rotor pre positioning phase also called alignment is necessary in order to make absolute the otherwise relative position information fed back by the quadrature encoder This alignment phase is performed only at first startup after any detected microcontroller fault event or reset Refer to Section 4 3 for a deeper description of this procedure After the rotor pre positioning is performed if ENCODER is uncommented the variables containing the target value of the Iq and Id stator current components respectively PID_Torque_InitStructure Reference and PID_Flux_InitStructure Reference are ini
95. ind of configuration utilized for the speed position feedback ENCODER or VIEW_ENCODER_FEEDBACK uncommented in the stm32f10x_MCconfh configuration file As already stated in the previous paragraph a rotor pre positioning phase also called alignment is necessary in this case Refer to Section 4 3 for a deeper description of this procedure After the rotor pre positioning is performed if ENCODER is uncommented the variables containing the target values of the Iq and Id current components PID_Torque_InitStructure Reference and PID_Flux_InitStructure Reference respectively are driven by the torque and flux controller block and the main state machine switches from the Start to the Run state On the other hand if VIEW_ENCODER_FEEDBACK is uncommented the ramp up strategy related to the sensorless operation starts just after the end of the pre positioning HALL_SENSORS is uncommented in the stm32f10x_MCconfh configuration file The PID_Torque_InitStructure Reference software variable is driven by the flux and torque controller block from the moment the start command is given The software variable containing the electrical rotor angle is also initialized based on the digital value of the three Hall sensor outputs at that moment Finally the main state machine switches from the Start to the Run state NO_SPEED_SENSORS is uncommented in stm32f10x_MCcont h In case of sensorless motor driving a particular ramp up is necessary in order t
96. ing either cheap shunt resistors or market available isolated current Hall sensors ICS e the rotor speed position feedback block allows the system to handle Hall sensor or incremental encoder signals in order to correctly acquire the rotor angular velocity or position Moreover this firmware library provides sensorless detection of rotor speed position as described in Section 2 2 e the PlID controller blocks implement proportional integral and derivative feedback controllers current regulation e the Clarke Park Reverse Park amp Circle limitation blocks implement the mathematical transformations required by FOC UM0492 Introduction to the sensorless FOC of PM motors Figure 5 FOC algorithm structure torque control Power converter Figure 6 Speed control loop Power converter The igs and igs Current components can be selected to perform electromagnetic torque and flux control On the other hand they can be driven to implement a speed control loop as depicted in Figure 6 via the torque and flux controller block which implements a PID controller for speed regulation and a flux weakening strategy In that case also a means of detecting rotor speed is needed Ti 15 121 Introduction to the sensorless FOC of PM motors UM0492 2 1 1 16 121 PM motor structures Mainly there are two different PM motor constructions available a Inthe first one drawing a in Figure 7
97. lectual property contained therein or considered as a warranty covering the use in any manner whatsoever of such third party products or services or any intellectual property contained therein UNLESS OTHERWISE SET FORTH IN ST S TERMS AND CONDITIONS OF SALE ST DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY WITH RESPECT TO THE USE AND OR SALE OF ST PRODUCTS INCLUDING WITHOUT LIMITATION IMPLIED WARRANTIES OF MERCHANTABILITY FITNESS FOR A PARTICULAR PURPOSE AND THEIR EQUIVALENTS UNDER THE LAWS OF ANY JURISDICTION OR INFRINGEMENT OF ANY PATENT COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT UNLESS EXPRESSLY APPROVED IN WRITING BY AN AUTHORIZED ST REPRESENTATIVE ST PRODUCTS ARE NOT RECOMMENDED AUTHORIZED OR WARRANTED FOR USE IN MILITARY AIR CRAFT SPACE LIFE SAVING OR LIFE SUSTAINING APPLICATIONS NOR IN PRODUCTS OR SYSTEMS WHERE FAILURE OR MALFUNCTION MAY RESULT IN PERSONAL INJURY DEATH OR SEVERE PROPERTY OR ENVIRONMENTAL DAMAGE ST PRODUCTS WHICH ARE NOT SPECIFIED AS AUTOMOTIVE GRADE MAY ONLY BE USED IN AUTOMOTIVE APPLICATIONS AT USER S OWN RISK Resale of ST products with provisions different from the statements and or technical features set forth in this document shall immediately void any warranty granted by ST for the ST product or service described herein and shall not create or extend in any manner whatsoever any liability of ST ST and the ST logo are trademarks or registered trademarks of ST in various countries Information in this
98. ler direction 5 6 3 Electrical angle extrapolation implementation As shown in Figure 50 the speed measurement is not the only task performed in TIMx_IRQHandler Beside the speed measurement the high to low transition of the XORed signal also gives the possibility of synchronizing the software variable containing the present electrical angle In fact as can be seen in Figure 57 any Hall sensor transition gives very precise information about rotor position iy 87 121 Library functions UM0492 5 7 88 121 Figure 51 Hall sensor output transitions 3 sensors 120 ai14847 For this reason in the proposed solution the electrical angle is synchronized every time an IC occurs with an angle depending on the present state on the Hall sensor output spinning direction and PHASE_SHIFT see also Section 3 10 for indications on how to measure it Furthermore the utilization of the FOC algorithm implies the need for a good and constant rotor position accuracy including between two consecutive falling edges of the XORed signal which occurs each 120 electrical degrees For this reason it is clearly necessary to somehow interpolate rotor electrical angle information For this purpose the latest available speed measurement in dpp format is added to the present electrical angle software variable value any time the FOC algorithm is executed Sensorless speed position detection Mc_State_Observer
99. m 20W Z1 STGF7NC60HD R2 BCSS7B R4 BRAKE_GPIO_ PORT AAA BRAKE _GPIO DIN i BCS47B Note that the size of the resistor in terms of both resistance and sustainable power should be carefully dimensioned When using the PMSM FOC library in conjunction with the STM3210B MCKIT note that the pin 23 of the MC connector J7 that carries the signal for brake implementation is positioned close to the wrapping area Note on debugging tools The third party JTAG interface should always be isolated from the application using the MB535 JTAG opto isolation board it provides protection for both the JTAG interface and the PC connected to it 37 121 Running the demo program UM0492 Caution During a breakpoint when using the JTAG interface for the firmware development the motor control cell clock circuitry should always be disabled if enabled a permanent DC current may flow in the motor because the PWM outputs are enabled which could cause permanent damage to the power stage and or motor A dedicated bit in the DBGMCU_SR register the DBG_TIM1_STOP bit must be set to 1 See Figure 25 In the main c module the DBG gt CR DBG_TIM1_STOP instruction performs the above described task Figure 29 DBG_TIM1_STOP bit in TIM1 control register extract from STM32 reference manual DBGMCU_CR Address 0xE0042004 Only 32 bit access supported POR Reset 0x00000000 not reset by system reset
100. m UM0492 Note 3 12 Caution 36 121 3 Run the motor in sensorless mode processing the sensor feedback The transformation blocks and the speed regulator will utilize the rotor position and speed information coming from the sensorless algorithm observed value The real measured information coming from the real sensors are processed for comparison with the observed ones With this aim it is necessary to Uncomment NO_SPEED_SENSORS and a definition between VIEW_HALL_FEEDBACK and VIEW_ENCODER_FEEDBACK depending on the sensor you are using Fill startup section of MC_State_Observer_param h and relax at first statistic parameters if motor cannot start with different current and frequency startup parameters Error on speed fdbck or Start up failed faults Be aware that due to the different speed resolution accuracy a different setup for speed PID could be necessary Tune startup parameters performing different ramp up trials and if required further tune observer and PLL gains 4 Congratulations The motor can now run in pure sensorless mode Depending on the requirements of the used debug feature on the expected code size and CPU load you could Comment VIEW_HALL_FEEDBACK or VIEW_ENCODER_FEEDBACK depending on which was not commented in stm32f10x_MCconf h Comment OBSERVER_GAIN_TUNING in stm32f10x_MCconf h Comment DAC_FUNCT
101. machine that is fed with a three phase system of positive sequence In that case to properly work the software library expects the Hall sensor signal transitions to be in the sequence shown in Figure 26 for both 60 and 120 displaced Hall sensors For these reasons it is suggested to follow the instructions given below when connecting a Hall sensor equipped PM motor to your board 1 Turn the rotor by hand in the direction assumed to be positive and look at the B emf induced on the three motor phases For this purpose if the real neutral point is not available it can be reconstructed by means of three resistors for instance 2 Connect the motor phases to the hardware respecting the positive sequence Let Phase A Phase B and Phase C be the motor phases driven by TIM1_CH1 TIM1_CH2 and TIM1_CHS respectively e g when using the MB459 board a positive sequence of the motor phases could be connected to J5 2 1 and 3 3 Turnthe rotor by hand in the direction assumed to be positive look at the three Hall sensor outputs H1 H2 and H3 and connect them to the selected timer on channels 1 2 and 3 respectively making sure that the sequence shown in Figure 26 is respected 4 Measure the delay in electrical degrees between the maximum of the B emf induced on Phase A and the first rising edge of signal H1 Enter it in the MC_hall_param h header file HALL_PHASE_SHIFT For your convenience an example with HALL _PHASE_SHIFT e
102. menus for setting control parameters when in Torque Control mode The parameter highlighted in red color is the one that can be set and its value can be modified by acting on the joystick key Moving the joystick up down selects the active control mode in the example shown in Figure 12 it is Torque control Once the motor Start command has been issued by pressing the JOY or KEY key this parameter is no longer accessible It becomes accessible again when the motor is stopped Figure 12 LCD screen for Torque control settings STMZ2 Motor Control PMSM FOC ver 1 0 Sensorless Demo Torque control mode Target Measured Iq 023000 00000 Id 09334 00000 Speed 00000 Move tlChange From the previous screen Figure 12 if the joystick is moved to the right the Target lq current component becomes highlighted in red This parameter can now be modified by moving the joystick up down Once the motor Start command has been issued Target Iq can be changed in runtime while the measured Iq current component is shown in the Measured field Figure 13 LCD screen for Target lq settings STMZ2 Motor Control PMSM FOC ver 1 0 Sensorless Demo Torque control mode Target Measured Iq 0323000 00000 Id 09334 00000 Speed 00000 Move tlChange From the previous screen Figure 13 if the joystick is moved to the right the Target Id current component becomes highlighted in red This parameter can now be modif
103. mics of the system can be assigned using a proper value of Kj See Figure 60 114 121 ki UM0492 Additional information Figure 60 Block diagram of closed loop system after pole zero cancellation Target R current Target 1 1 current gt R B 1 Ss K AB ai14853 It is important to note that the K and Kp parameters used inside the PI algorithms will be scaled using the proper divider The KpDIV and K DIV dividers are defined in MC_PID_regulators c so the computed values of Kp and K must be multiplied by these factors Moreover the PI algorithm does not include the PI sampling time T in the computation of the integral part See the following formula t n n kif e t dt kT X e kT K Se kT 0 k 1 k 1 Since the integral part of the controller is computed as a sum of successive errors it is required to include T in the Kj computation So the final formula can be expressed as Kp LopeKpDIV AB x Rs K DIV l AB Vb DC R A AB Bus shunt op 3 3 Usually it is possible to set the bandwidth of the closed loop system to 1500 rad s to obtain a good trade off between dynamic response and sensitivity to the measurement noise The Agp measured for the MB459 is 2 57 It is then possible to compute the values of the parameters knowing the motor parameters Rg Ls VgygDC and Rehunt 115 121 Additional information UM0492 A 5 Note Note 116
104. mizing the workspace for your STM32F103xx derivative 12 1 4 1 Inkarm_xxx xcl1 file internal Flash or RAM based project 12 1 4 2 Extended linker file setting 0 00 c eee eee eee 13 2 Introduction to the sensorless FOC of PM motors 14 2 1 Introduction to the PM synchronous motor FOC drive 14 2 1 1 PM motor structures 0 00 cece eee eee 16 2 1 2 PMSM field oriented control FOC fundamental equations 17 2 2 Introduction to sensorless rotor position speed feedback 18 2 3 Introduction to flux weakening control 0 0000 e eee eee 20 3 Running the demo program 00 cee eee eee eee 22 3 1 Torque control mode lt eis x ceeed we Vee RS eek wR Sd EEO R ESOS 22 3 2 Speed control mode 2 0 eee eee 24 3 3 Currents and speed regulator tuning 00 eee eee 25 3 4 Observer and PLL gain tuning 00 0 c eee eee eee 27 3 5 DAC functionality n n anaa oa ees oad ewe edad cena eae cele ds x2 27 3 6 Power stage feedbacks 000 cece eens 28 3 7 Fault messages acne stirs de taeee Ouse nrnna 28 3 7 1 OVENCUITENE s ceriti T Aenea Mae a he eG E ade a ete des 29 3 7 2 Overheating 00sec tenes 29 3 7 3 Bus overvoltage 0 cee ee ee 29 3 7 4 Bus undervoltage yirs issis 0 eee eee 30 3 7 5 Startup failed 0 0 0 0 tees 30 3 7 6 Error on speed fdock 0000 30 3 8 S
105. msidivj Stop 1 940 V 1 6100 V 1 6000 V E EEG LeCroy 1112712007 4 40 37 AM z 117 121 Additional information UM0492 A 6 118 121 Figure 62 KP 8000 and KI 1000 File Vertical Timebase Trigger Display Cursors Measure Math Analysis Utilities Help cal z Measure Pirms C4 P2 top C4 P3 rms C4 P4cfreq C2 PS PE value 7 0 mY 2 mY status 7 oR 200 mvidiv 200 m idiv 2 00 msidivf Stop 1 940 1 6100 Y 1 6000 V 20 0kS5 1 0MS sfEdge Positive LeCroy 11 27 2007 4 39 58 AM A priori determination of state observer gains In order to speed up the sensorless system development the user can follow the procedure described in this appendix to calculate the initial values of the state observer gains K1 and K2 Furthermore thanks to the implemented progressive system development described in Section 3 11 it is possible to get the best possible tuning for K1 and K2 The computation of the initial values of K1 and K2 is based on the placement of the state observer eigenvalues The required motor parameters are r motor winding resistance Ls motor winding inductance T sampling time of the sensorless algorithm which coincides with FOC and stator currents sampling as discussed in Section 4 2 The motor model eigenvalues could be calculated as iT S e 1 7 S e 1 The observer eigenvalues are placed with ey Ctobs F e _ 2 20bs F UM0492 Additional info
106. n has been lost Hall sensor timeout or that speed is decreasing very sharply in case of sensorless operation the quality of the soeed measurement expressed in terms of the distribution around the mean value is not good This typically means that either the observer is not properly tuned or that the speed is so low that a good observation of the induced B emf is not possible e g rotor is locked Refer also to Section 4 5 3 UM0492 Running the demo program 3 8 Caution 3 8 1 Setting up the system when using ICS sensors The default configuration provides for the use of three shunt resistors and no speed sensor This section gives you information about how to provide the STM32F103xx with ICS feedback signals and to properly customize the firmware When using two ICS for stator current reading you must ensure that the conditioned sensors output signal range is compatible with the STM32F103xx supply voltage In order for the implemented FOC algorithm to work properly it is necessary to ensure that the software implementation of the stm32 10x_svpwm_ics module and the hardware connections of the two ICS are consistent As illustrated in Figure 24 the two ICS must act as transducers on motor phase currents coming out of the inverter legs driven by STM32F103xx PWM signals PWM1 Phase A and PWM2 Phase B In particular the current coming out of inverter Phase A must be read by an ICS whose
107. n list shown below define MAX_MODULATION_100_PER_CENT up to 11 4 kHz PWM frequency define MAX_MODULATION_99_PER_CENT up to 11 8 kHz PWM frequency define MAX_MODULATION_98_PER_CENT up to 12 2 kHz PWM frequency define MAX MODULATION_97_PER_CENT up to 12 9 kHz PWM frequency define MAX MODULATION_96_PER_CENT up to 14 4 kHz PWM frequency define MAX MODULATION_95_PER_CENT up to 14 8 kHz PWM frequency define MAX MODULATION_94_PER_CENT up to 15 2 kHz PWM frequency define MAX MODULATION_93_PER_CENT up to 16 7 kHz PWM frequency define MAX MODULATION_92_PER_CENT up to 17 1 kHz PWM frequency define MAX MODULATION_91_PER_CENT up to 17 5 kHz PWM frequency For information on selecting the PWM switching frequency you will find advice in Section A 2 on page 111 To determine the max modulation index corresponding to the PWM switching frequency refer to Table 2 on page 64 Encoder feedback processing stm32 10x_encoder module List of available functions and interrupt service routines The following is a list of available functions as listed in the stm32 10x_encoder h header file e ENC _Init on page 79 ENC_Get_Electrical_Angle on page 79 ENC_Get_Mechanical_Angle on page 79 ENC_ResetEncoder on page 80 ENC_Clear_Speed_Buffer on page 80 ENC_Get_Mechanical_Speed on page 80 ENC_Calc_Average
108. nally TIMx input capture on negative edge of the XORed signal and overflow Update event interrupts are enabled RCC_APB1PeriphClockCmd RCC_APB2PeriphClockCmad GPIO_Structlnit GPIO_Init TIM_Delnit TIM_TimeBaseStructlnit TIM_TimeBaselnit TIM_ICStructinit TIM_ICInit TIM_PrescalerConfig TIM_InternalClockConfig TIM_SelectHallSensor TIM_SelectInputTrigger TIM_SelectSlaveMode TIM_UpdateRequestConfig NVIC_Init TIM_ClearFlag TIM_ITConfig TIM _SetCounter TIM_Cmd STM32F103xx reference manual section Interfacing with Hall sensors HALL GetRotorFreq Synopsis Description Functions called Input Returns See also s16 HALL_GetRotorFreg void This routine computes the rotor electrical frequency in dpp format starting from the array storing the latest period measurements and according to the following 16 formula a GST 3 SAMPLING_FREQ captured value prescaler value where CKTIM is the timer peripheral clock and SAMPLING_FREQ is the sampling rate of the FOC algorithm Be aware that speed is assumed to be zero if either the prescaler is equal to the maximum or a timeout occurred Please refer to Section 5 6 2 for more detailed explanation of the operating principle utilized for speed measuring GetAvrgHallPeriod or GetLastHallPeriod None It returns the electrical frequency in dpp unit Format is s16 Appendix A 7 Speed formats for more information about dpp unit UM0492 Library functions
109. nd Kd coefficient calculation as shown below UM0492 Library functions z Settings for min frequency define Freq Min 10 1 Hz mechanical define Ki_Fmin 1000 Frequency min coefficient settings define Kp_Fmin 2000 define Kd_Fmin 3000 Settings for intermediate frequency 1 define F_1 50 5 Hz mechanical define Ki_F_1 2000 Intermediate frequency 1 coefficient settings define Kp_F_1 1000 define Kd_F_1 2500 Settings for intermediate frequency 2 define F_2 200 20 Hz mechanical define Ki_F_2 1000 Intermediate frequency 2 coefficient settings define Kp_F_2 750 define Kd_F_2 1200 Settings for max frequency define Freq Max 500 50 Hz mechanical define Ki_Fmax 500 Frequency max coefficient settings define Kp_Fmax 500 define Kd_Fmax 500 Once the motor is running integer proportional and derivative coefficients are computed following a linear curve between F_min and F_1 F_1 and F_2 F_2 and F_max see Figure 54 Note that F_min F_1 F_2 F_max are mechanical frequencies with 0 1 Hz resolution for example F_1 1234 means F_1 123 4Hz Figure 54 Linear curve for coefficient computation Ki Kp Kd KiSF 2 KpUPs2 Kuh E I ee ee ei es a ee Ki_Fmax Kp_Fmax Kd_Fmax Ki F1 Kp F1 Kda Fi f Ki_Fmin Kp_Fmin Kd_Fmin ii Rotor mechanical F_min Fi F2 F_max frequency 99 121 Library functions UM0492 5 9 5 9 1 100
110. ng Therefore a startup procedure has been implemented in order to spin the motor when starting from standstill a rotating stator flux is generated by a three phase symmetrical current thus causing the rotor to follow The startup procedure has assumedly ended successfully when the observation of the back emfs becomes reliable according to the parameters explained in Section 4 5 3 and the main state machine switches from Start to Run otherwise a timeout occurs in that case the main state machine switches from Start to Fault The parameters described in this section are used to adapt the startup to the application by customizing the amplitude and frequency profiles see Figure 31 of the three phase current system UM0492 Getting started with the library Figure 31 Startup current system frequency and amplitude profile A Current frequency FINAL_START_UP_SPEED gt FREQ START_UP_DURATION t a Current amplitude FINAL_I _ START_UP s FIRST_ _ START_UP I H _______ p gt _START_UP_DURATION FREQ _START_UP_DURATION t define FREQ_START_UP_DURATION In milliseconds defines the overall time allowed for startup define FINAL_START_UP_SPEED In RPM defines the speed of the rotating stator flux and hence of the rotor at the end of the overall time allowed for startup this parameter sets the slope of the frequency linear ramp up define FIR
111. ng the update repetition rate based on the PWM frequency for 3 shunt resistor configuration on page 111 for details its value is 8 bit long Power board protection thresholds e define NTC_THRESHOLD C e define NTC_HYSTERIS_C These two values expressed in C are used to set the operating temperature range of the power devices measured at heat sink when the software library is used with the MB459 board In particular if the measured temperature exceeds NTC_THRESHOLD_C a fault event is generated that is kept as long as the measured temperature remains below NTC_THRESHOLD_C NTC_HYSTERESIS_C e define OVERVOLTAGE THRESHOLD_V e define UNDERVOLTAGE_THRESHOLD_V These two values expressed in volt set the minimum and maximum acceptable bus DC voltage when the software library is utilized with the MB459 board If the bus voltage exceeds OVERVOLTAGE_THRESHOLD_V or is below UNDERVOLTAGE_THRESHOLD_V a fault event is generated that is kept as long as the bus voltage remains outside the allowed range Speed loop sampling time define PID SPEED SAMPLING _TIME The speed regulation loop frequency is selected by assigning one of the defines below define PID SPEED SAMPLING _500us 0 min 500us define PID_SPEED_SAMPLING_1ms 1 define PID_SPEED _SAMPLING_2ms 3 4 1 500u
112. nput This function transforms stator voltage vg and vg belonging to a rotating frame synchronous with the rotor to a stationary reference frame so as to obtain v and ve See Section 5 4 2 for details Stator voltages Vg and Vas in q1 15 format as members of the variable Volt_Input which is a structure of type Volt_Components Stator voltages v and Vg in q1 15 format as members of a structure of type Volt_Components mul_q15_q15_q31 UM0492 Library functions Rev_Park Circle Limitation Synopsis void RevPark_Circle_Limitation void Description After the two new values vg and v of the stator voltage producing flux and torque components of the stator current have been independently computed by flux and torque PIDs it is necessary to saturate the magnitude of the resulting vector equal to v v before passing them to the Rev_Park function The purpose of this routine is to perform the saturation Refer to Section 5 4 3 Circle limitation on page 76 for more detailed information Input None Returns None Note The limitation of the stator voltage vector must be done in accordance with the PWM frequency as shown in Table 2 PWM frequency vs maximum duty cycle relationship on page 64 Functions called None Trig_Functions Synopsis Trig_ Components Trig_Functions s16 hAngle Description This function returns trigonometric cosine and sine functions of the input angle Input An angle in s16 format corre
113. ns Functions called TB Wait Synopsis Description Input Returns Functions called Caution void TB_Init void The purpose of this function is to initialize the STM32 system tick timer to generate an interrupt every 500 us thus providing a general purpose timebase None None SysTick_CLKSourceConfig SysTick_SetReload SysTick_CounterCmd NVIC_SystemHandlerPriorityConfig SysTick_ITConfig void TB_Wait u16 time This function produces a programmable delay equal to variable time multiplied by 500us Unsigned 16 bit None None This routine exits only after the programmed delay has elapsed Meanwhile the code execution remains frozen in a waiting loop Care should be taken when this routine is called at main interrupt level a call from an interrupt routine with a higher priority than the timebase interrupt will freeze code execution TB Set Delay 500us TB_Set_DisplayDelay 500us TB Set_StartUp_ Timeout TB Set _DebounceDelay 500us Synopsis Description Input Returns Functions called void TB_Set_Delay_500us u16 void TB_Set_DisplayDelay_500us u16 void TB_Set_StartUp_Timeout u16 void TB_Set_DebounceDelay_500us These functions are used to respectively update the values of the hTimebase_500us hTimebase_display_500us hStart_Up_TimeBase_500us and hKey_debounce_500us variables They are used to maintain the main state machine in Fault state to set the refresh rate of the LCD the Startup timeo
114. o make the rotor move and the sensorless algorithm converge to the actual rotor position A more detailed description of the ramp up procedure is described in Section 4 5 3 3 Currents and speed regulator tuning As already exposed in Section 2 1 the Iq and Id currents regulation is achieved by mean of two PID controllers where the derivative action can be optionally disabled by uncommenting the definition of Id_Iq_DIFFERENTIAL_TERM_ENABLED in stm32f10x_MCcont h Next figures show the two LCD menus allowing the real time tuning of the proportional integral and in case it is present derivative gains Figure 17 shows the screen used to select either of the torque PID coefficients whereas Figure 18 shows the screen used to select either of the flux PID coefficients From both screen either of the P or D when present coefficient can be selected highlighted in red by moving the joystick to the right left Then each value can be changed incremented or decremented by pressing the joystick up down Ti 25 121 Running the demo program UM0492 26 121 Figure 17 LCD screen for setting the P term of torque PID STM32 Motor Control PMSM FOC ver 1 90 Torque P I 023000 01000 Target 03000 Iq Measured 00000 Move tlChange Figure 18 LCD screen for setting the P term of flux PID STM32 Motor Control PMSM FOC ver 1 90 Flux P I 0230
115. observed currents ips as computed at step k by function STO_Calc_Rotor_Angle None Observed currents ip s16 format In the demo program this function is used only to display the variable of interest through DAC functionality if enabled in stm32f10xMCConf h STO _Get_wBemf_alfa_est Synopsis Description Input Returns Note s16 STO_Get_wBemf_alfa_est void This function returns the observed back emf e as computed at step k by function STO_Calc_Rotor_Angle None Observed back emf e s16 format In the demo program this function is used only to display the variable of interest through DAC functionality if enabled in stm32f10xMCConf h 93 121 Library functions UM0492 5 8 5 8 1 94 121 STO _Get_wBemf_beta_est Synopsis s16 STO _Get_wBemf_beta_est void Description This function returns the back emf eg as computed at step k by function STO_Calc_Rotor_Angle Input None Returns Observed back emf eg S16 format Note In the demo program this function is used only to display the variable of interest through DAC functionality if enabled in stm32f10xMCConf h STO Gains Init Synopsis void STO_Gains_Init StateObserver_Const StateObserver_ConstStruct Description This function is called by STO_StateObserverInterface_Init to initialize the Sensorless algorithm according to default parameters defined in MC_State_Observer_param h Returns None Note None STO Gains Update Synopsis
116. on detection algorithm is declared not reliable None Boolean TRUE if the observer provides reliable data This function relies on STO_IsSpeed_Reliable UM0492 Library functions STO Start_Up Synopsis Description Input Returns Note void STO_Start_Up void This function implements a startup procedure to be used to spin the motor when starting from standstill it has to be called with the same stator currents sampling periodicity As a result according to parameters set in MC_State_Observer_param h see Section 4 5 2 a rotating stator flux is generated by a three phase symmetrical current thus causing the rotor to follow During these operations the STO_Calc_Rotor_Angle function is called if the reliability of the observer is within the limits fixed in MC_State_Observer_param h see Section 4 5 3 the main state machine is allowed to switch to Run None None None STO Get_wlIalfa_est Synopsis Description Input Returns Note s16 STO _Get_wlalfa_est void This function returns the observed currents i as computed at step k by function STO_Calc_Rotor_Angle None Observed currents i S16 format In the demo program this function is used only to display the variable of interest through DAC functionality if enabled in stm32f10xMCConf h STO _Get_wIbeta_est Synopsis Description Input Returns Note s16 STO_Get_wlbeta_est void This function returns the
117. opsis Description Input Returns Functions called Park Synopsis Description Input Returns Functions called Rev_Park Synopsis Description Input Returns Functions called Curr_Components Clarke Curr_Components Curr_Input This function transforms stator currents ias and ip which are directed along axes each displaced by 120 degrees into currents i and ig in a stationary a B reference frame of axes are directed along paths orthogonal to each other See Section 5 4 2 for the details Stator currents ias and ips in q1 15 format as members of the variable Curr_Input which is a structure of type Curr_Components Stator currents i and ig in q1 15 format as members of a structure of type Curr_Components mul_qi5_q15_q31 Curr_Components Park Curr_Components Curr_Input s16 Theta The purpose of this function is to transform stator currents i and ig which belong to a stationary a B reference frame to a reference frame synchronous with the rotor and properly oriented so as to obtain igs and igs See Section 5 4 2 for details Stator currents i and ig in q1 15 format as members of the variable Curr_Input which is a structure of type Curr_Components rotor angle 6 65536 pulses per revolution Stator currents ig and igs in q1 15 format as members of a structure of type Curr_Components mul_qi5_q15_q31 Trig Functions Volt_Components Rev_Park Volt_Components Volt_I
118. or user interfacing None Bus voltage in volt units MCL_Compute_Temp Synopsis Description Input Returns u8 MCL_Compute_Temp void This function computes the power stage heat sink temperature in Celsius degrees only for MB459 board In the proposed firmware library this function is utilized for user interfacing None An integer representing a temperature value expressed in Celsius degrees 105 121 Library functions UM0492 106 121 MCL Calc BusVolt Synopsis void MCL_Calc_BusVolt void Description This function performs the averaging of the latest BUS_AV_ARRAY_SIZE temperature measurement by means of the following formula Xav K X y K 1 BUS_AV_ARRAY_SIZE 1 X K BUS_AV_ARRAY_SIZE where Xay K is the average at step K and X K the latest measurement at step K Input None Returns None the averaged value is written into a module private variable MCL Calc BusVolt Synopsis s16 MCL_Get_BusVolt void Description This function simply exports the averaged value of the bus voltage private variable Input None Returns Bus voltage in digits MCL Init Arrays Synopsis void MCL_Init_Arrays void Description This function initializes the averaged values of both voltage and temperature To be called after a MCU reset Input None Returns None MCL Brake Init Synopsis void MCL_Brake_Init void Description Declared and defined only if the brake resistor feature has be
119. ort Fortunately as discussed in Section 5 1 3 to reconstruct the currents flowing through a generic three phase load it is sufficient to simultaneously sample only two out of three currents the third one being computed from the relation given in Section 5 1 3 Thus depending on the space vector sector the A D conversion of voltage V will be performed only on the two phases where the duty cycles applied to the low side switches are the highest In particular by looking at Figure 33 you can deduct that in sectors 1 and 6 the voltage on the Phase A shunt resistor can be discarded likewise in sectors 2 and 3 for Phase B and finally in sectors 4 and 5 for Phase C Moreover in order to properly synchronize the two stator current reading A D conversions it is necessary to distinguish between the different situations that can occur depending on PWM frequency and applied duty cycles The explanations below refer to space vector sector 4 They can be applied in the same manner to the other sectors UM0492 Library functions Case 1 Duty cycle applied to Phase A low side switch is larger than DT Ty Where e DT is dead time Ty is the duration of the noise induced on the shunt resistor voltage of a phase by the commutation of a switch belonging to another phase Tsis the sampling time of the STM32F103xx A D converter the following consideration is made under the hypothesis that Ts lt DT Ty Refer to the STM32F103xx ref
120. output has to be sent to the analog channel specified by the PHASE_A_ADC_CHANNEL parameter in MC_pwm_ics_prm h Likewise the current coming out of inverter Phase B must be read by the other ICS and its output has to be sent to the analog channel specified by the PHASE_B_ ADC CHANNEL parameter in MC_pwm_ics_prm h About the positive current direction convention a positive half wave on PHASE_X_ADC_CHANNEL is expected corresponding to a positive half wave on the current coming out of the related inverter leg see direction of in Figure 24 Figure 24 ICS hardware connections Phase A Phase B Selecting PHASE_A_ADC_CHANNEL and PHASE_B_ADC_CHANNEL Default settings for PHASE_A_ADC_ CHANNEL and PHASE_B_ADC_CHANNEL are respectively ADC_CHANNEL11 and ADC_CHANNEL12 You can change the default settings if the hardware requires it by editing the Current reading parameters section of the MC_pwm_ics_prm h file As an example in order to convert Phase X X A B current feedback on ADC channel 0 the related parameters must be edited as shown below 31 121 Running the demo program UM0492 3 9 32 121 define PHASE X ADC CHANNEL ADC_Channel_0 define PHASE X GPIO PORT GPIOA define PHASE X GPIO PIN GPIO_Pin_0 Setting up the system when using an encoder Quadrature incremental encoders are widely used to read the rotor position of electric machines As the name implies incremen
121. pacitors For more detailed information on brake resistor management see also Section 3 12 29 121 Running the demo program UM0492 3 7 4 3 7 5 3 7 6 30 121 Bus undervoltage The bus voltage is below 20 V DC This threshold is specified in the MC_Control_Param h header file by the UNDERVOLTAGE_THRESHOLD_V parameter Refer to STM3210B MCKIT user manual for details Startup failed Available only when NO_SPEED_SENSORS is not commended It signals that no startup output condition was detected during motor ramp up FREQ_START_UP_DURATION in MC_State_Observer_param h See also Section 4 5 2 Error on speed fdbck An error on the speed position feedback was noticed Depending on the utilized kind of feedback this could mean that if an encoder is being used the measured speed was out of the allowed range MINIMUM_MECHANICAL_SPEED_RPM MAXIMUM_MECHANICAL_SPEED_RPM fora consecutive number of times equal to or higher than MAXIMUM_ERROR_NUMBER all these parameters can be found in MC_encoder_param h For instance this could mean that the encoder connection was lost See also Section 4 3 in case of a Hall sensors configuration the timer utilized for interfacing with the three Hall effect sensors overflowed for HALL_MAX OVERFLOWS MC_hall_param h consecutive times as mentioned in Section 4 4 and explained in Section 5 6 This usually indicates that informatio
122. plification network gain shunt current reading cell B11 51 121 Getting started with the library UM0492 52 121 As a result of data processing the following information can be obtained from the spreadsheet define MOTOR_RATED_SPEED_RPM This is the motor rated speed according to the definition above it is the highest speed for which the torque is maximum The user can copy it from cell B15 the mechanical limit of the motor must not be exceeded in any case define EXT_SPEED_RANGE This is the ratio between the maximum and the rated rotor speed cell B16 in the spreadsheet define ID_REFERENCE For each rotor speed beyond the rated one this is the table that defines the most suitable value of the reference current component ig digit This table has to be copied from cell B17 define IQ SATURATION For each rotor speed beyond the rated one this is the table that defines the saturation value for the current component ig digit This table has to be copied from cell B18 UM0492 Library functions 5 5 1 5 1 1 Library functions Functions are described in the format given below Synopsis Lists the prototype declarations Description Describes the functions specifically with a brief explanation of how they are executed Input Gives the format and units Returns Gives the value returned by the function including when an input valu
123. pling end HEE Space available for sampling ai14836b Case 2 DT Ty Ts 2 lt ADuty lt Dr Ty and ADuty ap lt Dr TRtTs With the increase in modulation index ADuty can have values smaller than Dy Ty Sampling synchronized with the counter overflow could be impossible In this case the two currents can still be sampled between the two Phase A low side commutations but only after the counter overflow Consider that in order to avoid the acquisition of the noise induced on the phase B current feedback by phase A switch commutations it is required to wait for the noise to be over Ty See Figure 38 Figure 38 DT Ty Ts 2 lt ADuty lt D7 Ty and ADutyap lt Dy TptTs ADuty ADuty Low side phase A 2 High side phase A A Low side phase B r5 High side phase B i oy hil 5 f i Current feedback phase s a A e Sampling start i i i i i i Ts Current feedback phaseC Wt A DnA an Rna e Sampling end m Space available for sampling lt rad ai14837b Case 3 ADutyA lt DT Ty Ts 2 and ADuty a B gt DT TR Ts In this case it is no more possible to sample the currents during Phase A low side switch on Anyway the two currents can be sampled between Phase B low side switch on and 62 121 yy UM0492 Library functions Phase A high side switch off The choice was therefore made to sample the currents Ts us before of phase A high side swit
124. qual to 90 C is illustrated in Figure 27 Figure 27 Determination of Hall electrical phase shift Y HI L H2 120 TI H3 J Lat a H2 60 H3 gt HALL_PHASE_SHIFT 34 121 Ti UM0492 Running the demo program 3 11 Progressive sensorless system development In order to simplify sensorless development a process was defined when designing the presented firmware library This process is a 4 step path that takes the user in a very short time to the final achievement of running the motor without either speed or position feedback sensors The defined process is based on the hypothesis that within the development stage of design the user can count on speed position rotor information coming from either the Hall sensors or the encoder If not possible or not convenient in the application it is anyway sufficient to mechanically couple the shaft of the motor to the one of the motor mounting sensors e g the BLDC motor provided in the STM3210B MCKIT This feedback information is actually necessary mainly for the purpose of comparing the rotor position information observed by the sensorless algorithm to the real one In this way the sensorless algorithm can be fine tuned As mentioned above the path consists of 4 steps listed and described below 1 Run the motor in a pure sensor configuration To this purpose you should Tune the lq and Id current loop regulator
125. rmation A 7 A 8 Typically by rule of the thumb the user can set f 4 Then the initial values of K1 and K2 could be calculated as fs C1obs 20bs 2 K ops 2obs T L S K L 1 10bs amp 20bs 4 obs 20bs 2 T Finally K1 and K2 could be used to fill in MC_State_Observer_param h see Section 4 5 1 Speed formats Two speed formats are commonly utilized in the PMSM FOC firmware library e 0 1 Hz this format is normally utilized by the speed regulators and by the highest layer of the software for user interfacing for instance digit per PWM dpp the dpp format expresses the speed as the variation of the electrical angle expressed in s16 format within a PWM period This format is particularly convenient since the rotor angular position can be easily determined by accumulating the rotor speed information every time the control loop is executed for example during PWM update interrupt service routine Providing that 2n OxXFFFF so that angle roll overs do not need to be managed the frequency with 0 1 Hz unit can easily be converted into dpp format using the following formula 65536 pp 1 Hz T0 SAMPUNG FREG Where SAMPLING_FREQ is the FOC sampling rate automatically computed in pre compilation phase starting from REP_RATE and PWM_FREQ References e 1 PC Krause O Wasynczuk S D Sudhoff Analysis of Electric Machinery and Drive Systems Wiley IEEE Press 2002
126. ry functions UM0492 104 121 MCL _ChkPowerStage Synopsis Description Functions called Input Returns MCL ClearFault Synopsis Description Functions called Input Returns See also MCL SetFault Synopsis Description Functions called Input Returns See also void MCL_ChkPowerStage void This function performs checks of the power stage working conditions only for temperature and bus voltage and when required generates a FAULT event MCL_Chk_OverTemp MCL_Chk_BusVolt MCL_SetFault None None bool MCL_ClearFault void This function checks if the cause of the fault event is over In the positive and if the Key button has been pressed the related flag is cleared and a TRUE is returned Otherwise a FALSE is returned TB_Delay_IsElapsed MCL_Chk_BusVolt MCL_Chk_OverTemp GPIO_ReadInputDataBit KEYS_ExportbKey None TRUE if all the fault flags are cleared and the Key button has been pressed by the user FALSE otherwise Section 3 7 Fault messages void MCL_SetFault u16 On occurrence of a fault event this function puts the main state machine in Fault state and disables the motor control outputs of Advanced Control Timer TIM1 PWM timer TB_Set_Delay_500us TIM1_CtrlPWMOutputs SVPWM_3ShuntAdvCurrentReading in case of three shunt current reading configuration Source of fault event as defined in MC_const h None Section 3 7 Fault messages UM0492 Li
127. s 2ms define PID_SPEED_ SAMPLING _5ms 9 define PID_SPEED_SAMPLING_10ms 19 define PID_SPEED_ SAMPLING _20ms 39 define PID_SPEED_SAMPLING_127ms 255 max 255 1 500us 127ms UM0492 Getting started with the library Speed P D controller init values define PID_SPEED_REFERENCE_RPM Define here in rpm the mechanical rotor speed setpoint at startup in closed loop mode define PID_SPEED_KP_DEFAULT The proportional constant of the speed loop regulation signed 16 bit value adjustable from 0 to 32767 define PID_SPEED_KI_DEFAULT The integral constant of the speed loop regulation signed 16 bit value adjustable from 0 to 32767 define PID_SPEED_KD_DEFAULT The derivative constant of the speed loop regulation signed 16 bit value adjustable from 0 to 32767 Quadrature current PID controller init values See Appendix A 4 which illustrates the method to be followed for computing initial PI constants starting from motor parameters and required control bandwidth define PID_TORQUE_REFERENCE The torque Iq reference value in torque control mode at startup signed 16 bit value define PID_TORQUE_KP_DEFAULT The proportional constant of the torque loop regulation signed 16 bit value adjustable from 0 to 32767 define PID_TORQUE_KI_DEFAULT The integral constant of the torque loop regulation signed
128. s fed with a three phase system of positive sequence Because of this and because of how the encoder output signals are wired to the microcontroller input pins it is possible to have a sign discrepancy between the real rolling direction and the direction that is read To avoid this kind of reading error apply the following procedure ky UM0492 Running the demo program 3 10 1 Turn the rotor by hand in the direction assumed to be positive and look at the B emf induced on the three motor phases For this purpose a neutral point may need to be reconstructed with three resistors if the real one is not available 2 Connect the motor phases to the hardware respecting the positive sequence for instance when using the MB459 board a positive sequence of the motor phases may be connected to J5 2 1 and 3 3 Run the firmware in encoder configuration and turn by hand the rotor in the direction assumed to be positive If the measured speed shown on the LCD is positive the connection is correct otherwise it can be corrected by simply swapping and rewiring the encoder output signals If this is not practical a software setting may be modified instead in the stm32f 10x_encoder c file replace the code line 164 TIM_ICPolarity_Rising TIM_ICPolarity_Rising by TIM_ICPolarity_Rising TIM_ICPolarity_Falling Setting up the system when using Hall effect sensors Hall effect sensors ar
129. s16 PID_Speed_regulator PID_SpeedTYPEDEF PID_Speed s16 speed Description The purpose of this function is to compute the proportional integral and derivative terms if enabled see SPEED_DIFFERENTIAL_TERM_ENABLED in Section 4 1 on page 39 for the speed regulation Input PID_SpeedTYPDEF see MC_type h for structure declaration signed 16 bits Output signed 16 bits Functions called None Caution Default values for the PID speed regulation are declared and can be modified in the MC_Control_Param h file see Section 4 2 on page 41 See also Figure 58 on page 114 shows the PID block diagram PID Reset_Integral_terms Synopsis void PID_Reset_Integral_terms void Description The purpose of this function is to reset all the integral terms of the torque flux and speed PID regulators PID Speed_Coefficients_update Synopsis void PID_Speed_coefficients_update s16 motor_speed Description This function automatically computes the proportional integral and derivative gain for the speed PID regulator according to the actual motor speed The computation is done following a linear curve based on 4 set points See Section 5 8 4 on page 98 for more information Functions called None Caution Default values for the four set points are declared and can be modified in the MC_Control_Param h file see Section 4 2 on page 41 PID Integral _Speed_update Synopsis void PID_Integral_Speed_update s32 value Description The purpose of
130. spondence with radians is illustrated in Figure 43 Returns Cosine and sine of the input angle in s16 format see Figure 44 as members of a structure of the Trig_ Components type Functions called None ky 73 121 Library functions UM0492 74 121 Figure 43 Radians versus s16 angle s16 32767 32768 2T ai14842 angle radians Figure 44 s16 versus sine and cosine cosine ine 32767 angle s16 ai14843 lt UM0492 Library functions 5 4 2 Detailed explanation about reference frame transformations PM synchronous motors show very complex and time varying voltage equations By making a change of variables that refers stator quantities to a frame of reference synchronous with the rotor it is possible to reduce the complexity of these equations This strategy is often referred to as the Reference Frame theory 1 Supposing fax fox fox are three phase instantaneous quantities directed along axis each displaced by 120 degrees where x can be replaced with s or r to treat stator or rotor quantities see Figure 45 supposing fqx fax fox are their transformations directed along paths orthogonal to each other the equations of transformation to a reference frame rotating at an arbitrary angular velocity can be expressed as 2r 2r 6 6 6 cos cos 3 cos 0 5 2 fadox fgx 3 sin sin 0 22 sin o 5 fox 1 Cx 1 1 L 2 2
131. t The IAR config folder contains 3 files e Inkarm_flash xcl C Inkarm_ram xcl These files are used as an extended command linker file and contain linker options Memory areas start address size and other parameters are declared here Refer to the Cortex M3 Technical Reference Manual for more information The default extended linker file used in the standard library to configure the device for internal Flash based resident firmware is Inkarm_flash xc1 An extract of this file showing the definitions of heap and stack size is provided below Depending on the project requirements it may be necessary to manually edit the segment sizes Code memory in FLASH DROMSTART 0x8000000 DROMEND 0x801FFFF Data in RAM DRAMSTART 0x20000000 DRAMEND 0x20004FFF RRR RR RRRERKEKK RRR EEK KKK KK RRR RK KEK RRR KERR KKK KKK KKERKKKKKKKKEKKK KKK KKK KKKKKKKK Stack and heap segments RRR RRRRRREKRKKK RRR RRR KKK KKK KR REKK EK RRR RE KR KKK KKK KKERKEKKEKKKKKEKKK KKK KKK KKKKKKKK D_CSTACK_SIZE 800 D_HEAP_STZE 400 Z DATA CSTACK _CSTACK_SIZE RAMSTART RAMEND Z DATA HEAP _HEAP_STZE RAMSTART RAMEND Memory size modifications might also be necessary according to the MCU specifications Default settings are done for a 128 KB embedded Flash memory If you use a different device you must edit the Inkarm_flash xcl file as explained in Section 1 4 2 ky UM
132. t routine UM0492 Library functions 5 10 Power stage check up MC_MotorControl_Layer module 5 10 1 List of available functions The following is a list of available functions as listed in the MC_MotorControl_Layer h header file e MCL_Init on page 103 MCL_ChkPowerStage on page 104 MCL_ClearFault on page 104 MCL_SetFault on page 104 MCL_Chk_OverTemp on page 105 MCL_Chk_BusVolt on page 105 MCL_Compute_BusVolt on page 105 MCL_Compute_Temp on page 105 MCL_Calc_BusVolt on page 106 MCL_Calc_BusVolt on page 106 MCL_Init_Arrays on page 106 MCL_Brake_Init on page 106 MCL_Set_Brake_On on page 107 MCL_Set_Brake_Off on page 107 MCL_Init Synopsis void MCL_Init void Description This function implements the motor control initializations to be performed at each motor start up it affects PID regulators current reading calibration speed sensors and high side driver boot capacitors initializations Functions called PID_Reset_integral_terms ENC_Clear_Speed_Buffer or HALL _InitHallMeasure and HALL_Init_Electrical_Angle or STO_Init depending on the speed feedback configured TB_Set_StartUp_ Timeout TIM1_CtrlPWMOutputs TB_StartUp_Timeout_IsElapsed SVPWM_3ShuntCurrentReadingCalibration or SVPWM_lIcsCurrentReadingCalibration depending on the current feedback configuration SVPWM_3ShuntCalcDutyCycles or SVPWM_lIcsCalcDutyCycles depending on the current feedback configuration Input None Returns None Ti 103 121 Libra
133. tal encoders actually read angular displacements with respect to an initial position if that position is known then the rotor absolute angle is known too For this reason it is always necessary when processing the encoder feedback ENCODER or VIEW_ENCODER_FEEDBACK definitions not commented in stm32f10x_MCconf h to perform a rotor prepositioning before the first startup after any fault event or microcontroller reset Quadrature encoders have two output signals represented in Figure 25 as TI1 and T12 With these and with the STM32F103xx standard timer in encoder interface mode it is possible to get information about the rolling direction Figure 25 Encoder output signals counter operation forward jitter backward jitter forward I Bree yg Pe ice WS Gig EP pS EN ees ga Oe AY ae TI2 ee peAa LI p COUNTER up down up zU opipa pi In addition the rotor angular velocity can be easily calculated as a time derivative of the angular position To set up the PMSM FOC software library for use with an incremental encoder simply modify the stm32 10x_MCconf h and MC_encoder_param h header files according to the indications given in Section 4 1 and Section 4 3 respectively However some extra care should be taken concerning what is considered to be the positive rolling direction this software library assumes that the positive rolling direction is the rolling direction of a machine that i
134. tartInjectedConvCmd ADC_GeiFlagStatus ADC_GetInjectedConversion Value SVPWM_IcsInjectedConvConfig Motor Control library SVPWM_IcsInjectedConvConfig SVPWM_IcsGetPhaseCurrentValues Synopsis Description Input Returns Note Functions called Curr_Components SVPWM_IcsGetPhaseCurrentValues void This function computes current values of Phase A and Phase B in q1 15 format from the values acquired from the A D converter None Curr_Components type variable In order to have a q1 15 format for the current values the digital value corresponding to the offset must be subtracted when reading phase current A D converted values Thus the function must be called after SVPWM_lIcsCurrentReadingCalibration None UM0492 Library functions SVPWM_lIcsCalcDutyCycles Synopsis void SVPWM_IcsCalcDutyCycles Volt_Components Stat_Volt_Input Description After execution of the FOC algorithm new stator voltages component Va and Vg are computed The purpose of this function is to calculate exactly the three duty cycles to be applied to the three inverter legs starting from the values of these voltage components Refer to Section 5 1 2 for details about the theoretical approach of SVPWM and its implementation Input Va and Vg Returns None Note None Functions called None 5 2 2 Current sampling in isolated current sensor topology and integrating general purpose A D conversions The three currents l4 lo and Ig flowin
135. te s16 STO_Get_Speed void It returns the rotor electrical speed as STO_Calc_Speed calculated by averaging the buffered values of observed speed None Observed rotor electrical speed dpp format See Section A 7 about the speed format STO _Get_Speed_Hz Synopsis Description Input Returns Note s16 STO_Get_Speed_Hz void It returns the rotor mechanical speed None Observed rotor mechanical speed Hz 10 This function relies on STO_Get_Speed STO IsSpeed_ Reliable Synopsis Description Input Returns Note bool STO_IsSpeed_Reliable void This routine indicates if the information provided by the sensorless algorithm is reliable To do so it checks the module private flag managed by STO_Calc_Speed A FALSE statement is an indication of a malfunctioning in the rotor position reconstruction due for example to an improper choice of the observer and or PLL gains or to a locked rotor condition None Boolean TRUE if the observer provides reliable data None STO Check Speed Reliability Synopsis Description Input Returns Note bool STO_Check_Speed_Reliability void This routine indicates if the information provided by the sensorless algorithm has remained reliable over time It should be called with the same speed sampling time periodicity STO_lIsSpeed_Reliable is called if that function returns FALSE for RELIABILITY_HYSTERESYS MC_State_Observer_param h times then the rotor speed positi
136. that can be set and its value can be modified by acting on the joystick key From the menu screen shown in Figure 15 it is possible to switch from Torque control to Speed control operations and vice versa by moving the joystick up down while the motor is stopped Figure 15 Speed control main settings STM32 Motor Control PMSM FOC ver 1 90 Sensorless Demo Speed control mode Target Measured 01500 rpm 00000 Move tlChange From the menu screen shown in Figure 16 moving the joystick to the right selects the Target speed parameter highlighted in red Once selected the parameter can be incremented decremented by moving the joystick up down The motor can then be started simply by pressing the joystick When the motor is on the target speed can still be modified Figure 16 LCD screen for setting Target speed STM32 Motor Control PMSM FOC ver 1 90 Sensorless Demo Speed control mode Target Measured 01500 rpm 00000 Move tlChange Like in the torque control mode the motor is started stopped by pressing the joystick or the KEY button Since in speed control mode the torque and flux parameters Target Iq and Target Id are the outputs of the Torque and flux controller they cannot be set directly The PID regulators can however be real time tuned as explained below UM0492 Running the demo program Different motor ramp up strategies are used in speed control mode depending on the k
137. that the phase voltages must be centered at 50 of duty cycle it follows that the values to be loaded into the PWM output compare registers are given respectively by Ti 2 X Z Sector I IV TimePhA z 5 TimePhB TimePhA Z TimePhC TimePhB X T T 2 Y Z Sector Il V TimePhA 4 TimePhB TimePhA Z TimePhC TimePhA Y Sector III VI TimePhA i T 2 Y X TimePhB TimePhC X TimePhC TimePhA Y Current sampling in three shunt topology and general purpose A D conversions The three currents l4 lo and Ig flowing through a three phase system follow the mathematical relation ly lo 1g 0 For this reason to reconstruct the currents flowing through a generic three phase load it is sufficient to sample only two out of the three currents while the third one can be computed by using the above relation The flexibility of the STM32F103xx A D converter makes it possible to synchronously sample the two A D conversions needed for reconstructing the current flowing through the motor The ADC can also be used to synchronize the current sampling point with the PWM output using the external triggering capability of the peripheral Owing to this current conversions can be performed at any given time during the PWM period To do this the ky UM0492 Library functions Figure 34 control algorithm uses the fourth PWM channel of TIM1 to synchronize the start of the conversions Injected conversions as described abo
138. the application layer Moreover it is possible to get rid of any speed sensor thanks to the sensorless algorithm for rotor position reconstruction A prerequisite for using this library is basic knowledge of C programming PM motor drives and power inverter hardware In depth know how of STM32F103xx functions is only required for customizing existing modules and for adding new ones for a complete application development Figure 1 shows the architecture of the firmware It uses the STM32F103xx standard library extensively but it also acts directly on hardware peripherals when optimizations in terms of execution speed or code size are required Figure 1 Firmware architecture Application layer Speed flux and STM32F103xx torque PIDs FOC dive user standard interface Speed Current library feedback feedback SVPWM STM32F103xx motor control library STM32F103xx peripherals ai14812 Rev 1 1 121 www st com Contents UM0492 Contents 1 Getting started with tools 00 eee eee 10 1 1 Working environment 22 scedeeee ade eee beebtadan nsdn beiades wed 10 1 2 Software 1001S nkepe an ee Ss OU Ree ERA ERE OY OE a 10 1 3 Library source COdG s o2 40s eee stb es GbE eee Ea oe eee ee esas 11 1 3 1 Download 4 0 5 acai ad dae a Raha da iaa a aaa aed es Rae ads 11 1 3 2 File SUUCIUIC sosiete ie hbe iva thawed adabse Heda ads 11 1 3 3 Starting the IAR toolchain 0 0 c eee 11 1 4 Custo
139. tialized with the values PID_TORQUE_REFERENCE and PID_FLUX_REFERENCE defined in the header file MC_Control_Param h the main state machine switches from the Start to the Run state On the other hand if VIEW_ENCODER_FEEDBACK is uncommented the ramp up strategy related to the sensorless operation starts just after the end of the pre positioning HALL_SENSORS is uncommented in the stm32f10x_MCconfh configuration file In this case no rotor pre positioning is performed and the PID_Torque_InitStructure Reference software variable is simply initialized with the PID_TORQUE_REFERENCE value defined in the MC_Control_Param h header file The software variable containing the electrical rotor angle is initialized based on the digital value of the three Hall sensor outputs and the main state machine switches from the Start to the Run state NO_SPEED_SENSORS is uncommented in stm32f10x_MCconf h In case of sensorless motor driving a particular ramp up is necessary in order to make the rotor move and the sensorless algorithm converge to the actual rotor position A deeper description of the ramp up procedure is described in Section 4 5 23 121 Running the demo program UM0492 3 2 24 121 Speed control mode Figure 15 and Figure 16 show two LCD menus used to set control parameters when in Speed control mode The parameter highlighted in red color is the one
140. ting CPU load related to FOC algorithm execution 110 A 2 Selecting the update repetition rate based on the PWM frequency for 3 shunt resistor configuration 0 005 111 A 3 Fixed point numerical representation 0000 eee eee 112 A 4 A priori determination of flux and torque current PI gains 113 A 5 Current regulators fine tuning 20 cee eee eee eee 116 4 121 UM0492 Contents A 6 A priori determination of state observer gains 4 118 A 7 SPSGUTONNAIG ce svegeiies sad Soe us Weened Meese bee eee eee ears 119 AS References si0cs cece sive Keb eea ede ewes Bia we eee en ayeee ee 119 Revision history cca dared chennai ewe ee ee Ae hee hee 120 ky 5 121 List of tables UM0492 List of tables Table 1 Sector identification 0 0 eee eee 58 Table 2 PWM frequency vs maximum duty cycle relationship 0 00 e eee eee 64 Table 3 PWM frequency vs maximum duty cycle relationship 0 00 eee eee 77 Table 4 System performance when using STM3210B MCKIT 0000 ee eee ee eee 112 Table 5 Document revision history 0 ects 120 6 121 Ti UM0492 List of figures List of figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 Figure 15 Figure 16 Figure 17 Figure 18 Figure 1
141. tor parameters i W4 gt W3 gt w gt W gt w current limit circle inverter motor rating With reference to Figure 11 itis possible to state that the motor thermal rating and power converter capability define a current limit circle every current vector inside this circle is allowed on a thermal basis On the other hand the DC bus voltage motor parameters and electrical frequency wdefine a voltage limit circle every current vector inside this circle could exist on an electrical basis Therefore for each frequency and hence for each rotor speed the current vectors allowed are those that belong to the intersection of the current limit circle and the specific voltage limit circle Each rotor speed determines a pair ig ig Max f a the whole set of pairs draws the trajectory of flux weakening operations With reference to Figure 11 it is possible to recognize as the motor rated speed and the red curve as the flux weakening and maximum torque per ampere trajectory By providing motor parameters see PMSM_Flux_Weakening xls spreadsheet included in CD ROM it is possible to precalculate the aforementioned current vector trajectory and insert the result in the proper parameter header file of the firmware library see Section 4 6 iy 21 121 Running the demo program UM0492 3 3 1 22 121 Running the demo program Torque control mode Figure 12 Figure 13 and Figure 14 show a few LCD
142. ual to 32767 signed 16 bits max If the current is measured by using shunt resistors then MAX_CURRENT where A 2 Rshunt v A is the gain of the amplifying network A 2 57 on the MB459B kit board Rsnunt is the shunt resistance in Ohms define BUS_ADC_CONV_RATIO Defines the ratio between ADC input voltage and corresponding DC bus voltage define K1 K1 signed 32 bit value is an element of the gain vector of the implemented state observer as described in Section 2 2 An a priori determination of K1 can be made using the formulas given in Section A 5 When the motor is running this initial value can then be tuned using the LCD interface and evaluating the results In that case the value of K1 read on the display is 10 times smaller 47 121 Getting started with the library UM0492 Note 4 5 2 48 121 define K2 K2 signed 32 bit value is an element of the gain vector of the implemented state observer as described in Section 2 2 An a priori determination of K2 can be made using the formulas given in Section A 5 When the motor is running this initial value can then be tuned using the LCD interface and evaluating the results In that case the value of K2 read on the display is 100 times smaller define PLL_KP_GAIN The default formula provides an a priori determination of the phase detector gain of the PLL Nonetheless this signed 16 bit value can be tuned
143. ue according to the formula o gt u VARIANCE_THRESHOLD where cand u are the variance and the mean value of the observed speed respectively for instance a VARIANCE_THRESHOLD of 0 0625 leads to a percentage of the mean value equal to 25 The lower the VARIANCE_THRESHOLD parameter the more strict and hence the higher the safety level of this fault detection algorithm and vice versa define RELIABILITY_HYSTERESIS This unsigned 8 bit parameter defines the number of consecutive times the speed measurement variance should be found higher than the VARIANCE_THRESHOLD threshold before the rotor speed position detection algorithm is declared nonreliable In this case the main state machine switches from Run to Fault see Section 3 7 about the fault messages and Section 5 7 about the functions that implement this method define NB_CONSECUTIVE_TESTS Defines the number of consecutive times the soeed measurement variance should be found lower than the VARIANCE_THRESHOLD threshold before the startup procedure is declared to have successfully completed In this case the main state machine switches from Start to Run see Section 3 7 about the fault messages Permanent magnet synchronous motor parameters MC _PMSM motor _param h The MC_PMSM_motor_param his to be filled with the motor parameters Three different sections can be
144. ut and to filter the user key bouncing Unsigned 16 bits None None 101 121 Library functions UM0492 TB StartUp _Timeout_IsElapsed TB Delay IsElapsed TB DisplayDelay_IsElapsed TB_DebounceDelay_ IsElapsed Synopsis Description Input Returns Functions called SysTickHandler Synopsis Description Input Returns Functions called Note 102 121 bool TB_StartUp_Timeout_IsElapsed void bool TB_Delay_IsElapsed void bool TB_DisplayDelay_IsElapsed void bool TB_DebounceDelay_IsElapsed void These functions return TRUE if the related delay is elapsed FALSE otherwise None Boolean None void SysTickHandler void This is the System Tick timer interrupt routine It is executed every 500us as determined by TB_Init and is used to refresh various variables used mainly as counters for example PID sampling time Moreover if FLUX_TORQUE_PIDs_TUNING is uncommented in stm32f10xMCConf it controls the current component reference iq to generate a square wave of defined period see Section 4 1 and Appendix A 5 None None FOC_CalcFluxTorqueRef PID_Speed_Coefficients_update If Encoder is used ENC_Calc_Average_Speed if using DAC ENC_Get_Mechanical_ Speed If Hall sensors are used if using DAC HALL_GetSpeed if using the sensorless algorithm STO_Calc_Speed STO_Check_Speed_Reliability MCL_SetFault STO_Obs_Gains_Update if using DAC STO_Get_Speed This is an interrup
145. ve are used for current reading purposes whereas regular conversions are reserved for the user As soon as the injected A D conversions for current reading purposes have completed bus voltage and temperature sensing are also simultaneously converted by the dual A D Figure 34 shows the synchronization strategy between the TIM1 PWM output and the ADC The A D converter peripheral is configured so that it is triggered by the rising edge of TIM1_CH4 PWM and ADC synchronization TIM1_CH1 TIM1_CH2 TIM1_CH3 TIM1_CH4 Counter overflow OCR 4 OCR 1 OCR 2 OCR 3 TIM1_CH1 TIM1_CH2 TIM1_CH3 TIM1_CH4 Sampling point before counter overflow Sampling point after counter overflow ai14835 In this way supposing that the sampling point must be set before the counter overflow that is when the TIM1 counter value matches the OCR4 register value during the upcounting the A D conversions for current sampling are started If the sampling point must be set after the counter overflow the PWM 4 output has to be inverted by modifying the CC4P bit in the TIM1_CCER register In so doing when the TIM1 counter matches the OCR4 register value during the downcounting the A D samplings are started After the first two simultaneous conversions other two simultaneous conversions are started one for the bus voltage and the other for the temperature sensing At the end of the second conversion the thre
146. void STO_Gains_Update StateObserver_GainsUpdate STO_GainsUpdateStruct Description This function is called by STO_Obs_Gains_Update to modify state observer and PLL gains Returns None Note None Currents and speed regulators MC_PID_ regulators module The MC_PID_regulators module contains all the functions required for implementing the necessary PID regulators for controlling currents los and Ipg and in case of Speed control mode motor speed List of available functions The following is a list of available functions in the MC_PID_regulators module PID_Init on page 95 PID_Flux_Regulator on page 95 PID_Torque_Regulator on page 95 PID_Speed_Regulator on page 96 PID_Reset_Integral_terms on page 96 PID_Speed_Coefficients_update on page 96 PID_Integral_Speed_update on page 96 UM0492 Library functions PID_Init Synopsis Description Input Functions called Note void PID_Init PID_Torque TYPEDEF PID_Torque PID_FluxTYPEDEF PID_Flux PID_SpeedTYPEDEF PID_Speed The purpose of this function is to initialize the PIDs for current and speed regulation For each a set of default values is loaded target speed or current proportional integral and derivative gains lower and upper limiting values for the output PID_TorqueTYPEDEF PID_FluxTYPEDEF PID_SpeedTYPEDEF see MC_type h for structure declarations None Default values for PID regulators are declared and can be modified in t
147. voltage constant K V krom RMS phase to phase in Volts Additional parameters for flux weakening operation This section has to be filled if depending on the application specifics the drive has to operate the motor beyond its rated speed here rated speed is considered to be the highest speed at which the motor can deliver maximum torque The mechanical limit of the motor must not be exceeded in any case Moreover if the motor is being operated beyond the rated speed it is mandatory to have a regenerative power converter or a brake resistance see Section 3 12 See Section 2 3 about the strategy implemented for flux weakening control define MOTOR_VOLTAGE_CONSTANT Defines the motor voltage constant Ke V krpm RMS phase to phase in Volts The editing of the following parameters requires that the annexed spreadsheet is filled with the information below e Mains AC voltage Volts RMS cell B e Motor rated current 0 to peak Amperes cell B2 These data must be matched with inverter current rating Motor winding inductance mH phase cell B3 Motor voltage constant K Volts RMS krpm phase to phase cell B4 Motor winding resistance Ohms cell B5 Number of magnetic pole pairs cell B6 Maximum operating speed required by the application in the flux weakening region cell B7 Maximum modulation index MMI cell B9 It is a function of the applied PWM frequency see Section 4 2 Shunt resistance Ohms cell B10 Am

UM0492 User manual

Contents

Download Pdf Manuals

Related Search

Related Contents