View AN1327_104839.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

AN1327 APPLICATION NOTE
BRUSHLESS DC MOTOR: SENSORLESS FUZZY CONTROL BY ST52X420
Author: G. Grasso, V. Marino
ABSTRACT
This application note describes the very low-cost implementation of a Brushless DC Motor actuator. The cost reduction is obtained by means of an 8-bit fuzzy microcontroller which implements both a Sensorless phase driving and a closed loop speed control. The application is designed for high-voltage motors suitable for 220Vac direct use. Keywords: BLDC, Sensorless, Permanent Magnet Rotor, Speed control, Main supply.
1. INTRODUCTION
In the last years the home appliance market is becoming a strong competitive field. The growing performances/costs ratio is leading the industry to new approaches in the motor control field in order to satisfy EMI/EMC laws, Quality and reliability standards, etc. The results of these constraining factors are: long life motors, maintenance-free motors and sensors removal requirements; high efficiency driving, and so on. Brushless DC motors well fit these requirements and are becoming widely used as small horsepower motors. Moreover, when the application does not require high dynamic performances in the handling of the load torque, a sensorless control of the BLDC is a suitable low-cost implementation. Adjustable-speed drives for home appliances systems, large fans, compressors, hydraulic pumps, ... etc are good examples of system in which the load does not change suddenly and then a sensorless approach could be enough.
2. BRUSHLESS MOTOR OVERVIEW
Brushless DC (BLDC) motors are referred to by many aliases: brushless permanent magnet, permanent magnet DC motors, permanent magnet synchronous motors, etc. This type of motors is generally driven (supplied) from an inverter, which converts a constant voltage to a 3-phase voltage with a frequency corresponding instantaneously to the rotor speed. Electronics replaces the function of commutator and energize the proper winding. The energized stator winding leads the rotor magnet and switches as the rotor aligns with the stator. The motor used in the current application is a three-phase motor, two pole pairs, Vrms=200V, Irms=300mA, used in a radial fan split for air conditioning system.
December 2000
1/19
AN1327 - APPLICATION NOTE
Figure 1. BLDC section- 3-phase, 1 pole pair for each phase
N
S
The following theoretical tips are not an outline on synchronous machines but a transposition of the concept in our application. For details on the BLDC motor principles the reader can refer to any book on electrical machines (see references). In synchronous motor drives, the stator is supplied with a set of balanced three-phase currents, whose frequency is f. If p is the number of the poles in the motor then:
f=
p s 4
(1)
where s (rad/s) is the flux synchronous speed or the rotor speed. In our case the stator has two pole pairs for each phase (p=4).This means that:
= 2f =
4 s = 2 s 2
In other words, for one turn of the rotor (360 mechanical), the inverter must supply two trapezoidal cycles for each phase. By using the per-phase equivalent model of the motor it is possible to make some considerations. In the diagram Rs is the phase stator winding resistance, La is the armature inductance and Lis is the winding leakage inductance.
2/19
AN1327 - APPLICATION NOTE
Figure 2.
Efa represents the Back-EMF (bemf). This is the voltage that the permanent magnet rotor induces in the stator coil when motion takes place. This EMF is always present during the normal driving of the motor and opposes a counter-force to the motion of the shaft that is proportional to the rotor speed. Main end of this application note is to detect the BEMB. The expression of the electromechanical torque is:
TEM = kt f Ia sin
where Kt is a constant, f is the field-flux, is called torque angle. represents the angle between the phase linked flux f and the relative stator current Ia. Now, if the electronic controller is able to supply the phase Ph1 in order to maintain =90 the fields decoupling takes place then the (2) can be simplified as:
TEM = KT IS
(3)
where KT is called motor torque constant and IS is the amplitude of the stator phase currents. In these specific conditions (i.e. stator flux in quadrature with the rotor flux) the equation (3) shows that the output torque is proportional only to the supplied current. This is the same case of a simple DC machine. Main task of this application note is to maintain the torque angle =90 without the use of any position sensor.
3/19
AN1327 - APPLICATION NOTE
3. INVERTER DRIVER TOPOLOGY
The widely used circuit to perform the trapezoidal wave driving of a DC brushless motor is the six-step inverter where each phase is driven by means of a couple of transistors. Fig.3 shows the basic operating principle of this drive. Figure 3. Drive basic operating principle
+Vbus
V/I
.
half rotor turn (360 electrical)
Vbatt
.
.
60 .
V An
.
.
ZCa
. t
T1 A Vbatt
T2
T3 Ph1 B C Ph2 Ph3
.
ZCa
.
IA
ZCa
t
n IC
IB
t
T4 -Vbus
T5
T6
t VI I
T1-T5
II
T1-T6
III
T2-T6
IV
T2-T4
V
T3-T4
VI
T3-T5
Sectors
States
N
T3-T5
The name "six-steps" arises from the finite time-steps in which the whole period 0/2 can be shared. Being our motor a two pole pairs, the electrical period 0/2 represents half turn of the rotor. During each time step, the current direction does not change whereas current amplitude can increase or decrease in the windings. This topology drives each winding for the 240electrical degrees, allowing to one phase at a time to be "floating" for 120. The floating state is the mandatory condition to read the BEMF produced by the rotor flux in the phase under evaluation. For instance in the third sector T2 and T6 are conducting; node 'A' is floating because switches T1 and T4 are open. If an oscilloscope is connected between node 'A' and 'n' during the III sector, an inducted voltage will be displayed. This is the voltage that a rotating permanent magnet induces on a fixed coil. Polarity inversion of this voltage means that the rotor flux inverts its direction from North-pole to South-pole or viceversa. This flux inversion (voltage zero-crossing) is a useful information to well know the rotor position. Like in the sensored applications we can take advantage of this position information in order to drive the next coils.
4/19
AN1327 - APPLICATION NOTE
4. HARDWARE DESCRIPTION
As already discussed, the main target of this application note is to reduce the overall cost of the system. For this reason, a simple voltage scaling in conjunction with the AD converter is used to detect the zero crossing occurrency. Fig. 4 shows the resistor partition used to limit the overvoltage on the AD inputs. Figure 4. BEMF sensing principle
Vph
.
T= Tm/2 V An
.
A B C
Ph1 Ph2 Ph3
ZCa
. . T/6 . . t
n
T/12
ZCa ZCb
.
V Bn
. III IV VI t
ZCa ZCb .
I .
T/12 + T/6 + T12
ZCb
To Analog-Digital Converter
Rsense
Only two phases are monitored in order to detect their zero-crossing. In fact, only two temporal information are enough to know position and speed of the rotor with enough accuracy. The AD converter will continuosly read the phase voltages and will produce an interrupt to the CPU when the Zero-Crossing is reached. At this point, a brief theoretical consideration on the acquired voltages must be done. As shown in fig. 5 the zero crossing event is referred to the phase voltages VAn, VBn (i.e. between phase terminal and center tap). The converted values of the scaled voltages are referred to the ground (line voltages). It is possible to transform the line voltages into the phase voltages using the center tap voltage and the translation equations. This is tipycally done with a DSP approach. In this application note a different method is used in order to detect the zero crossing event. For the six step inverter (discontinuous mode), it is easy to show that the center tap voltage is a constant value during the six sectors. Its level depends from the chopping time of the High Side switches when the PWM is used to control the speed. Assuming that the center tap voltage is constant, the AD can detect the ZC by comparing the actual converted value with a "middle-point" value. This middle value is the half of the BEMF ranges detected in the previous sectors.
5/19
AN1327 - APPLICATION NOTE
Figure 5. Line to center and line to neutral voltages
Vph
V An
. I II IV V VI t
ZCa
Max
VAN
.
middle-point level Min Vsat
. t
ZCa
The AD-subroutine (see later for details) is shared in several sub-tasks, one for each sector, in order to minimize the overall AD-interrupt time-task. For instance: during the I^ sector only the Max-value is sampled, filtered and divided to obtain 'middle' value; during the II^ sector the phaseB voltage is converted and compared with 'middle' value in order to flag the ZCb event; during the VI^ sector for the ZCa event. The 'Min' value is about zero volt being the dropout voltage on the low-side switches of the bridge. Another feature of the driver is the total current monitoring. A very little resistor value Rsense is used to sense the current level in the phases during each step of the inverter. This current measurement will be used to detect wrong motor operation and rotor asynchronism (pull-out torque).
6/19
AN1327 - APPLICATION NOTE
5 SOFTWARE DESCRIPTION
The following paragraphs describe the implementation of the Application note realized by using the ST52x420 microcontroller. The easy use of the microcontroller support tools allows to simply adjust and optimize the microcontroller code The main program presentation is carried out in a graphical way in order to describe in a user-friendly and intuitive manner the flow of the program. After a paragrah dedicated to the configuration of the ST52x420 peripherals (please refer to ST52x420 data sheet for further information), the flowchart of the main program and of the relative subroutines is provided.
5.1 ST52x420 Settings
In the following picture is shown the configuration of the peripherals of ST52x420 used in the current application (cfr: peripheral configuration chapter in ST52x420 data sheet). Figure 6. PWM configuration
PWM-Timer 0 is used in PWM mode, with a switching frequency of 19.6 KHz; T0_OUT is AND-ed with the three high-side driver signals in order to modulate the voltage on the motor and then establish its speed. PWM-Timer 1 and PWM-Timer 2 are used in Timer mode. The first timer is used to calculate the time between the zero-crossing of two phases (that is 1/3 of the electrical period T); because the prescaler is 2048 and the clock master is 20 MHz, the Timer 1 clock period is 102.4 s, then the counted period T/3 is given by 102.4 s*PWM_1_COUNT. Instead, the second timer is used to generate the driver signals. Because in each period T there are six switches command, Timer 2 is used to scan 6 temporal events. Of course these 6 events must be scanned at each T/6 and synchronized with the rotor position. For this purpose, Timer2 prescaler is set to divide by 1024 achieving a Timer2 clock input at 51,2s. By imposing: 51,2s x PWM2_COUNT=T/6 it is easy to see that it is enough to equal: PWM2_COUNT=PWM1_COUNT This is a simple method to obtain a period division (from T/3 to T/6) without the use of a DIV instruction.
7/19
AN1327 - APPLICATION NOTE
In other words, Timer2 counts at a speed rate that is the double of Timer1; then, with the same count value, Timer1 reads T/3 whether Timer2 imposes T/6. Figure 7. Timers configuration
The six driver signals are provided by the ST52x420 parallel port; these are supplied on the pins PA0-PA5, configured as output pins as shown in fig. 8. Figure 8. I/O configuration
8/19
AN1327 - APPLICATION NOTE
AD converter is configured with four channels, which are converted continually: Ain0 and Ain1 are used to detect ZCa and ZCb; Ain2 is used to convert the information coming from Rsense that is related with the overall maximum current; Ain3 is used to convert the voltage coming from the potentiometer that sets the target speed. Figure 9. AD Converter configuration
9/19
AN1327 - APPLICATION NOTE
5.2 Algorithm Description
The following figure summarises all the tasks implemented in the ST52x420 program. At the top is shown the main program while on the following parts are shown the AD and Timer2 interrupt service subroutines. The main program mainly consists of 2 different loops activated one after the other according to the variable mode: start-up and sensorless modes. Figure 10. Flow-chart .
Init var and peripherals
Start-up mode
mode
Synchronous sensorless mode
AD_converter Interrupt
Sensorless Mode
Yes
Read current
Find max phase
Yes
Max calculation
No Return from Interrupt
No Yes Find z. c. phase A No Find z. c. phase B Yes Detect z.c. phase B Read TIM1_COUNTER Reset TIM1 and start TIM2 Detect z.c. phase A and start TIM1 Return from Interrupt
No
Change phase and set flags Yes TIM2 Interrupt Send driver signals into PORT_A Return from Interrupt
Sensorless Mode No
Increase frequency and set duty-cycle
10/19
AN1327 - APPLICATION NOTE
a) Start-up mode
The main goal of this procedure is to lead the rotor from 0 rpm to almost 1000 rpm, in order to have some BEMF waveforms detectable by AD converter. During this operation mode the six driver signals are supplied with a set 6 of commands with duration of T/6. Progressively this starting period T is storten (stator flux speed increases) in order to force the rotor movement. For this puropose, Timer2 is used to synchronize the witches command from one secotr to another. As referred in the Timer2 flowchart,at each falling edge of Timer2_out, the variable 'phase' (that represents the sector) is incremented, in order to send into PORT_A the pattern related. Varying the PWM_2_COUNT value, the frequency of the driver signals is increased to reach the desired value; note that, in this mode, the duty-cycle of PWM0 is increased (phase voltage increases) according to the frequency value, by means of the fuzzy block present in the Timer2 interrupt routine.
b) Synchronous mode
After a programmed delay, the sensorless mode is started; in this mode a speed loop control is performed as shown in fig. 11, in order to reach the reference speed value; a fuzzy block, reading the speed error (the difference between the desired and the actual speed), allows to adjust the duty-cycle value (PWM), modulating the voltage on the motor and then adjusting the speed rotation. To read the rotation speed, a sensorless algorithm is implemented. By using the AD_converter to detect two subsequent zero-crossing events, a time measurement is done by means of Timer1. timer1 starts at the first zero-crossing and is stopped at the second event. This period represents, as preivously discussed, 1/3 of the electrical period T, and then, to have the time T/6, (used to commutate the driving signal in Timer2 interrupt routine), it is enough to load the value PWM_1_COUNT in PWM_2_COUNT. At each commutation, besides producing the six driving signals, some flags are set to establish the actions to perform (find max, find zero-crossing on phase A or on phase B). Finally, in the AD interrupt routine, the maximum current value Imax is sampled. If Imax overcomes a fixed current value the driver immediately turns off and the microcontroller is reset (thanks to the watchdog) to restart the program. Figure 11. Control topology
Speed_ref
Error Fuzzy Controller
PWM Motor Driver
11/19
AN1327 - APPLICATION NOTE
5.3 Main program window
Let us discuss about the main program; 'init_var' and 'int_enable' blocks are used to initialize global variables and interrupt mask (only ADC and Timer2 interrupts are enabled). The following blocks let ADC, PWM0 and Timer2 start. After starting, PWM2 will count for an initial period imposed from the variable named `period_start_up'. Nexy, this variables will be decreased (see Timer 2 Interrupt routine) in order to increment the rotor speed as discussed before. The following figure reports the main program in term of graphical programming. Figure 12. Main program
5.4 Fuzzy Control Window
By implementing the control topology described in fig. 11, the following figures (13 and 14) show the graphical code implementation. The block 'error_calc' performs error calculation as 'error=ref-period', where 'period' is the variable containing the value T/3 counted by Timer2. The `error' variable is sent to the Fuzzy engine. This `fuzzy block' produces the incremental value `dv' that is added 9with sign) to the current PWM0 value. the updated PWM0 duty-cycle is sent to the PWM peripheral by the block named `set_PWM0'. In the calculation blocks of fig. 13 and 14 the `IF' `THEN' operators are used to avoid overflow or underflow in the variables 'error' and 'duty' during the control. Figure 14 shows a classical integrative control described with the fuzzy linguistic aproach. Five Mbfs are chosen and five rules are set in order to implement a single input and output control block.
12/19
AN1327 - APPLICATION NOTE
Figure 13. Fuzzy Control
Figure 14. Fuzzy rules
13/19
AN1327 - APPLICATION NOTE
5.5 Timer 2 Interrupt window
The task of this subroutine is to generate the signals used to drive the switches T1-T6, either in start-up and synchronous mode. In the block 'change_sector' displayed in fig. 15, the variable 'phase' is incremented in order to switch to the next sector. The block 'send_driver_signals' is used to write a Boolean pattern on the PORT_A pins. Figure 15. Interrupt routine
This pattern is available in a look-up table stored in the ST52 x420 EPROM. Since the Timer2 subroutine is executed at each T/6 period then PORT_A will change its pattern at each T/6. The following figure shows the pattern for the driver topology previously discussed.
14/19
AN1327 - APPLICATION NOTE
Figure 16. Driver signal generation
PA0 PA1 PA2 PA3 PA4 PA5
0 0
1
0 0 0
1
0 0 0 0
0
0
0 0
1
0 0 0
1
0
1
0
1 1
0 0
1
0 0
1
0
1
0
1
1
T/6
20 17 33 34 10 12 Tabled values in the EPROM
The same look-up table is used to store some flags that are addressed during the variable's status change. These flags are used to indicate to the AD interrupt routine the task to be performed in that moment (find max, find ZCa or ZCb) accordingly to the related sector. Fig. 15 also shows a fuzzy routine (Fuzzy0) used only for the start-up phase. The fuzzy block implements a non linear relation between `period_start_up' and `duty' variables. As already stated, the stator flux speed is increased to let the rotor start.
5.6 ADC Interrupt window
Figure 17 shows the AD interrupt routine main window. Fig. 18 and 19 give more details on Folder1 and Folder2. In fact, this AD interrupt routine subroutine is shared into three sub-tasks, each one corresponding with the sectors I ('find_max'), II ('find_zcb') and VI ('find_zca'). During the first sector, when the phase A of the motor is connected to Vbus (T1 is on), the maximum value is continually acquired by Ain0 pin. The average value of these samples, divided by two (in the block 'max'), represents the value 'middle' to be compared to detect the zero-crossing on the phases A and B. Figure 17. AD interrupt routine
15/19
AN1327 - APPLICATION NOTE
A different task has to be performed in the sixth sector (find Zca) as shown in fig 18. Figure 18. Find Zero Crossing on phase A
The blocks `dec_delay_a' and `delay_a' implement a blanking time necessary to avoid the sampling of commutation spike. This spike arises from the coil opening in correspondence to the switching sector (See technical literature for further reference). After this blanking time, the value read on Ain0 is compared with the value 'middle' (block 'Jump1'). When sampled value on phase A voltage reaches the 'middle' value, the block 'strt_tim1' is executed, and Timer1 is started. Analogously, in the second sector, the zero-crossing on phase B is detected, and when this occurs, the block 'strt_tim2' is executed (fig. 19). In this block Timer1 is stopped, its value (PWM_1_COUNT) is read and Timer1 is reset. At the same time, Timer2 is preloaded with the read value of PWM_1_COUNT and then started.
16/19
AN1327 - APPLICATION NOTE
Figure 19.
The timers tasks described above allow to synchronize each BEMF zero-crossing with the rising/fallingedge of Timer2. Then the Timer2 edge events to control the commutation of the driver signals to be sent to Port A. Since Timer1 starts and stops on two BEMF events, its content also gives information about the rotor speed to implement the PID fuzzy controller.
17/19
AN1327 - APPLICATION NOTE
6. ELECTRICAL SCHEMATICS
Figure 20 shows the hardware implemented in the current application. The active devices are designed for a brushless motor of 70W and speed rate of 5.000 rpm. The user can easily modify the device values in order to meet his own specifications. Figure 20.
~ ~ +
V1
+VDC-POWER
+15Vcc 47nF/630V GP7NB60 2A500V 4 R1 220K Q3 C3 + C1 C4 T1 T1 2 100uF/450V T4 T4 SGND 1 Lin SD 3 1uF U2 Hin 1uF C7 Vboot HVG OUT LVG Pgnd Diag Cin 14 13 J2 12 9 8 2 5 6 R6 R5 10K 100 1 3 GP7NB60 Q4 L6386 OUT1 BLUE 1 OUT1 T5 T5 1 Lin 2 SD + 10uF R4 100 3 1 T2 T2 3 Hin 2 C2 U1
+15Vcc GP7NB60 C6 Vboot HVG OUT LVG Pgnd SGND Diag Cin 14 10uF 13 12 9 8 5 6 GP7NB60 OUT2 +
2 +15Vcc Q1 R2 100 3 Vcc T3 J1 1 ORANGE OUT2 T3 2 SD 3 Vboot HVG OUT LVG T6 2 T6 SGND Q5 R7 100 3 1 L6386 7 1 Lin Pgnd Diag Cin Hin C5 1 4 1uF U3 14 10uF 13 C8 + GP7NB60
2 Q2 R3 100 3 1
Vcc
Vcc
4
J3 12 9 8 2 5 6 GP7NB60 Q6 R8 100 3 1 OUT3 1 YELLOW OUT3
L6386 7
+15Vcc
7
R9 OUT2 OUT2 C9 R12 150K 10K 1 R15 R17 2 4.7V +15Vcc J4 2 1 220Vac TRANSFORMER 1:2 T1 3 U5 LM78L05ACZ VIN GND VOUT 1 OUT1 OUT1 + C16 C17 0.1uF 0.5A 50V 3 1000uF25V D1 4.7V 1 R16 10k 2 C11 22nF C12 22nF 9 10 11 12 15 16 17 18 22K R21 +5Vcc 2 C21 47k 180K R20 C22 0.1uF 1uF C23 C19 U6 Ain0/PB0 Ain1/PB1 Ain2/PB2 Ain3/PB3 Ain4/PB4 Ain5/PB5 Ain6/PB6 Ain7/PA7/PB7 Vdd VddA PA0/TORES PA1/T0OUTn PA2/T1OUTn PA3/T2OUTn PA4/T0STRT PA5/T0CLK PA6 PC0/INT PC1/T0OUT PC2/T1OUT PC3/T2OUT Oscout 1 Y1 1 3 2 ST52T420 3 20Mhz 9 8 10 U7C 7 74HC08 T3 2 0.1uF 28 13 25 24 23 22 21 20 19 5 6 7 8 2 +5Vcc 7 14 74HC08 C20 0.1uF +5Vcc 74HC08 +5Vcc 1 3 2 T4 T5 T6 7 T1 +5Vcc U7B 6 5 14 U7A C15 + 470uF16V C18 0.1uF 10K R14 phaseA +5Vcc 2 R10 150K Isense 3 10k C13 22nF C14 22nF R18 2.7k +5Vcc 8 D2 phaseB R11 R13 0.1ohm/1W C10 0.1uF 5 2.7k 330pF 14 + +5Vcc U4A TS902 2 470k
4-
~ ~ +
V2
+5Vcc 1 Speed Reference
R19
1 4 Reset 26 Test Vpp 27 14 3 Vss GndA Oscin
6.8uF
14 4
T2
7. CONCLUSIONS AND RESULTS
By using ST52x420 8-bit microcontroller it is easy to implement a low-cost approach to achieve a real time control. The brusless motor control described in this application note represents a good compromise between system costs and motor performances. The graphical programming environment reduces the development time also for non-expert programmers. The overall system has been implemented to control the speed of a fan in an air-conditioning system with good results.
18/19
3
AN1327 - APPLICATION NOTE
REFERENCES
[1] Paul C. Krause "Analysis of Electric Machinery" - McGraw-Hill [2] STMicroelectronics "Power products- Application manual" [3] Mohan, undeland, Robbins "Power Electronics: Converters, Applications and Design" - John Wiley & Sons [4] Yasuhiko Dote, Sakan Kinoshita "Brushless Servomotors - Fundamentals and Applications" Oxford Science Publications [5] FUZZYSTUDIOTM4.0- User Manual, STMicroelectronics 2000
Information furnished is believed to be accurate and reliable. However, STMicroelectronics assumes no responsibility for the consequences of use of such information nor for any infringement of patents or other rights of third parties which may result from its use. No license is granted by implication or otherwise under any patent or patent rights of STMicroelectronics. Specification mentioned in this publication are subject to change without notice. This publication supersedes and replaces all information previously supplied. STMicroelectronics products are not authorized for use as critical components in life support devices or systems without express written approval of STMicroelectronics. The ST logo is a trademark of STMicroelectronics (c) 2000 STMicroelectronics - All Rights Reserved FUZZYSTUDIOTM is a registered trademark of STMicroelectronics STMicroelectronics GROUP OF COMPANIES http://www.st.com Australia - Brazil - China - Finland - France - Germany - Hong Kong - India - Italy - Japan - Malaysia - Malta - MoroccoSingapore - Spain - Sweden - Switzerland - United Kingdom - U.S.A.
19/19

▲Up To Search▲

Price & Availability of AN1327

	To Download AN1327 Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .