A Low-Cost Sensorless Control for Reduced-Parts, Brushless DC Motor Drives

A. Halvaei Niasar, H. Moghbelli, A. Vahedi

Department of Engineering, University of Kashan, Kashan, Iran Department of Science and Mathematics, Texas A&M University at Qatar, Doha, Qatar
Center of Excellence for Power System Automation and Operation, Iran University of Science & Technology, Tehran, IRAN

Halvaei@ieee.org Abstract
Brushless DC (BLDC) motor is attracting much interest due to its good performance for many applications. Moreover, cost reducing of the drive is more attractive for low cost applications. This paper presents the design and implementation of a reduced parts BLDC motor drive. Part reducing is achieved by elimination of three Hall Effect position sensors and reducing the number of power switches to four switches. For current commutation, a low cost sensorless control based on line voltages is developed. Two second-order Butterworth low-pass filters with little phase delay are designed to eliminate of high frequency PWM and calculation of average terminal voltages. Proposed sensorless control doesn't need to any 30o or 90o phase shift that is prevalent in other sensorless methods. Moreover, to make the rectangular phase current waveforms, direct phase current control is used in which the currents of two phases A and B are controlled independently. The performance of the developed algorithms is verified via simulation and implementation. It is shown that the main source of the estimation error is drop voltage on the stator impendence in which at low speeds and for heavy loads it increases. Experimental results show that the developed sensorless control is effective and can be used for low cost applications.

II. ANALYSIS OF SENSORLESS-CONTROLLED FOUR-SWITCH, BLDC MOTOR DRIVE A. Four-Switch Inverter, BLDC Motor Drive Fig. 1 shows the three phase currents waveforms with the trapezoidal back EMF voltages of a BLDC motor. Typical mathematical model of a three-phase BLDC motor is as:
0 ia ea va R 0 0 ia L - M 0 d vb = 0 R 0 ib + 0 L - M 0 dt ib + eb 0 L - M 0 0 R 0 vc ic ic ec

(1)

I.

INTRODUCTION

The permanent magnet brushless DC (BLDC) motors increasingly are being used in computer, automotive, industrial and household products because of its high power density, compactness, high efficiency, low maintenance and ease of control. BLDC motor is inherently electronically controlled and requires six commutation points per cycle. Conventional BLDC motor drive is generally implemented via six-switch three-phase inverter and three Hall Effect position sensors that generate proper signals for current commutation [1]. On the other hand it is important to lower the manufacturing cost of the BLDC motor drive for many applications. Cost reduction of BLDC motor drive is accomplished by topological approach and the control approach. From a topology point of view, minimum number of switches and eliminating the mechanical sensors are required for the inverter circuit. In the control approach, using high performance processors, algorithms are designed and implemented to produce the desired characteristics [2]. In this paper, a low cost BLDC motor drive is introduced so that, both reducing the number of power switches and elimination the position Hall sensors are implemented.

BLDC motor needs quasi square current waveforms, which are synchronized with the back EMF to generate constant output torque. Also, at every mode only two phases are conducting and another phase is inactive. However, in the four-switch inverter, the generation of 120 conducting current profiles is inherently difficult. Hence, the direct phase current (DPC) control method is used [2]. Therefore, the currents of phase A and B in modes 2 and 5 are controlled independently and the current profiles are the same as the currents of a conventional six-switch inverter BLDC motor drive. Fig. 2 shows the overall system configuration of a four-switch inverter, three-phase inverter (FSTPI) BLDC motor drive.

Fig. 1. Signal waveforms of a BLDC motor

9 7 8 - 1 - 4 2 4 4 - 1 6 6 6 - 0 / 0 8 / $ 2 5 . 0 0 ' 2 0 0 8 I E E E6 6 2

Fig. 4. Determination of position signals from line voltages

Fig. 2. Basic configuration of FSTPI-BLDC motor drives via direct phase current control method

B. Novel Sensorless Control of FSTPI-BLDC Motor Drive Sensorless control of BLDC motor has been a research topic for the last two decades [3]. The sensorless techniques that utilize the back EMF voltage of the openphase are more attractive due to its simplicity and ease of implementation especially for low cost applications. There are many researches that have utilized the back EMF voltage or terminal voltages and based on the detection of zero crossing points. But most of them have been developed for conventional six-switch BLDC motor drive and actually, there aren!t any considerable researches on sensorless control of four-switch BLDC motor drive. Lately, in [4], from equality points of terminal voltages vao and vbo, six commutation points are determined. Accumulation error in this method may cause position error especially in transients. This paper presents a novel sensorless method for four-switch BLDC motor drive based on zero crossing points of stator line voltages. In a four-switch inverter topology, terminal C is connected to the middle point of DC bus (point o). Supposing the point o as reference, Fig. 3 shows the three line voltages waveforms named Vao, Vba and -Vbo. Zero crossing points (ZCPs) of these voltages lag 30o phase delay from ZCPs of phase back EMF voltages and so are coincide to the commutation instants. Therefore, with detecting the zero crossing points of three line voltages, six commutation points are obtained. Three line voltages are derived from terminal voltages Vao and Vbo easily. They have higher magnitude rather than back EMF voltages ("3 times plus drop voltage on the stator impedance).
V ao & ea 20 0 -20 0.05 20 0 -20 0.05 20 0 -20 0.05 0.1 time (sec) 0.15 0.2 Fig. 3. Stator line and phase back EMF voltages 0.1 0.15 0.2

Due to PWM control of the inverter, stator terminal voltages Vao and Vbo contains high frequency switching signals and so, detection the zero crossings of three line voltages is difficult. Hence, two low pass filters should be use to eliminate of high frequency harmonics and calculation the average of the line voltages. Fig. 4 shows the estimation of three position signals from zero crossing detection of line voltages. Estimation of the position signals can be carried out in hardware or software. Using DSP controller with embedded A/D inputs eliminates using extra hardware to detect zero crossing points and it can be implemented in software. However, implementation via hardware leads to lower cost and using notexpensive microcontrollers. Table I summarizes the relation between estimated position signals which is named virtual Hall signals and corresponding operation mode. III. SIMULATION In this section the performance of the developed sensorless algorithm is demonstrated via simulation in different conditions. Employed BLDC motor is a low speed motor that has got 16 poles and has been designed for electric bicycle applications. The motor parameters are given in the appendix. Two second-order Butterworth low pass filters with pass-band frequency of 700 rad/sec are used to calculate the average terminal voltages. Fig. 5 shows the drive performance at two speed 50 rpm and 200 rpm under 3 N.M load. The estimated mode as shown in Fig. 5 has a bit delay from real mode due to the estimation errors. Instantaneous terminal voltages and corresponding average values are shown that there are little delays between their zero crossings due to delay of filters. At two speeds 50 and 200 rpm the filter delay is 1o and 3.5o respectively. Virtual Hall signals are derived from three line voltages that are some delay from real ones because of drop voltage on the stator impedance. Moreover, phase currents waveforms are shown at two low and high speeds. Current glitches increase at low speeds because drop voltage on the stator impedance is considerable rather than the value of back EMF voltage.
TABLE I COMMUTATION LOGIC WITH RESPECT TO LINE VOLTAGES Line voltage Vba Vba -Vbo Vba Vao -Vbo Sign before ZCP + + + Sign after ZCP + + + Next mode 1 2 3 4 5 6 HcHbHa 100 101 001 011 010 110

V ba & eb

0.1

0.15

0.2

-V bo & ec

663

Mode
0.15 0.2 0.25 0.3 0.35 0.4
ao

Mode

5 0 0.1

5 0 0.02 0.04 0.06 0.08 0.1

20 0 -20 0.1 20 0 -20 0.1 0.15 0.2 0.25 0.3 0.35 0.4 10 0 -10 0.1 10 0 -10 0.15 10 0 -10 0.1 5 0 -5 0.1 5 0 -5 0.1 5 0 -5 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.15 0.2 0.25 0.3 0.35 0.4

20 0 -20 0.02 20 0 -20 0.02 0.04 0.06 0.08 0.1 20 0 -20 0.02 20 0 -20 0.02 20 0 -20 0.02 5 0 -5 0.02 5 0 -5 0.02 5 0 -5 0.02 0.04 0.06 0.08 0.1 0.04 0.06 0.08 0.1 0.04 0.06 0.08 0.1 0.04 0.06 0.08 0.1

[V]

ao

[V]

bo

ao

0.15

0.2

0.25

0.3

0.35

0.4

&H

ba

0.2

0.25

0.3

0.35

0.4

bo

-V

0.15

0.2

0.25

0.3

0.35

0.4

i [A]

ia [A]

-V

bo

&H

&H

ba

&H

ao

&H

&H

bo

[V]

[V]

0.15

0.2

0.25

0.3

0.35

0.4

0.04

0.06

0.08

0.1

i [A]

ib [A]

0.04

0.06

0.08

0.1

i [A]

0.15

0.2

0.25 0.3 time (Sec)

0.35

0.4

ic [A]

0.04

0.06 time (Sec)

0.08

0.1

(a) At speed of 50 rpm

(b) At speed of 200 rpm

Fig. 5. Simulation of the developed sensorless algorithm at different speeds

IV. EXPERIMENTAL RESULTS Developed sensorless algorithm has been demonstrated via a DSP controller TMS320LF2407A [5]. Fig. 6 shows the schematic of the implemented system that including: eZdsp board, switch drive circuits, current sensing circuits, zero crossing detection circuit, IGBT modules, BLDC motor and a 60 volt DC bus. DSP acts as controller to control the whole of the system. PWM pulses of DSP are sent to HCPL switch drivers for isolation and amplifying and then are applied to IGBT switches. The current of two phases A and B are measured using LA55-P current transducers for direct phase current control purposes. Virtual Hall signals are developed via hardware circuits. On this way as shown in Fig. 7, at first two terminal voltages are measured using LV25-P voltage transducers and then are filtered using active second-order Butterworth low pass filters. Next, three line voltages are derived from filtered terminal voltages. Finally, using Schmitt-trigger circuits, the zero crossings of derived line voltages are detected and virtual Hall signals are generated.

Designed software in DSP is based on a waiting loop interrupted by both the timer 2 unit (INT2) and the capture unit (INT4) of Event Manager A (EVA). At service routine of INT2, the currents are measured and regulated via hysteresis controllers. Switching frequency is set to 15 kHz. Service routine of INT4 detects the commutation instants as well as the operation mode of the motor.

Fig. 6. Schematic diagram of the developed sensorless-controlled, FSTPI-BLDC motor drive base on TMS320LF2407A DSP

664

Fig. 7. Developing the virtual Hall signals via hardware circuits

Fig. 8. Back EMF voltages of the employed BLDC motor

Fig. 8 shows the phase back EMF voltages waveforms and physical Hall position signal Ha. The back EMF waveforms are trapezoidal. In Fig. 9(a,b) the measured terminal voltages are shown at two low and high speed respectively. At high speed, due to effect of back EMF voltage of phase C, independent current controllers send the different commands to the power switches. It's the reason of PWM switching in Fig. 9(b). Terminal voltages are filtered to determine the average terminal voltages. Developed waveforms confirm the simulation results in Fig. 5. Moreover, three line voltages that are used for

sensorless algorithm are shown in Fig. 9(c,d). They are symmetrical and in 120o together and are used to make the virtual position signals. Estimated position signals are shown in Fig. 10 at low and high speeds (70 and 180 rpm) and under different load conditions. The low pass filters are well adjusted in which its phase delay at low and high speeds is 1o and 3o. It is a bit phase delay for many sensorless applications. Hence, as shown afterwards, the drop voltage on the stator is the main estimation error source.

(a) Terminal voltages at speed 70 rpm

(b) Terminal voltages at speed 250 rpm

(c) Three line voltages at speed 70 rpm

(d) Three line voltages at speed 250 rpm

Fig. 9. Terminal voltages and three line voltages waveforms at low and high speeds (voltage scale is 21v/v)

665

(a) At speed 70 rpm and no load condition

(b) At speed 180 rpm and 70% full load condition

(c) At speed 70 rpm and 70% full load condition

(d) At speed 180 rpm and 70% full load condition

Fig. 10. Estimated position signals rather than physical Hall signal Ha at different load and speed conditions

At low speeds, because the drop voltage on the stator impedance is considerable rather than the back EMF voltage magnitude, the estimation error increases. However, as long as the speed is increasing, the estimation error decreases. The most estimation error are belonged to the low speed and high load conditions that drop voltage on the stator is much high. Therefore this

sensorless algorithm is feasible for some applications that the high constant load is not applied to the motor at low speeds and load torque increases as long as the speed is increasing. Fig. 11 shows the phase currents at the startup of the BLDC motor and so no load condition.

(a) Motor startup

(b) No load

Fig. 11. Sensorless control of the motor in startup and no load conditions

666

(a) At speed 70 rpm

(b) At speed 200 rpm

Fig. 12. Current waveforms in the sensorless-controlled drive under 70% load condition at low and high speeds (current scale is 6A/v)

The motor in started in open-loop scheme as shown in Fig. 11(a). At first, the motor is aligned in mode 2 with exciting two phases A and B for 0.5 second. After that an open loop exciting is given for electrical cycle and then the developed sensorless algorithm is applied. The rotational loss of the employed motor is high that is seen in Fig. 11(b). Fig. 12 shows the sensorless control under 70% full load condition at two speeds 70 and 200 rpm. Current regulation is successful. Although, there are some current distortion due to effect of the back EMF voltage ec. The direct phase current control has not regulated the phase current perfectly for the employed motor and the drive needs to more adjusting. V. CONCLUSION The driving system of a reduced parts BLDC motor drive which can be used in low cost applications has been studied in this paper. The main benefits are increased system reliability and cost reduction of the overall system. Cost reduction is achieved via reducing number of power switches and related circuits such as power supplies and drivers and also elimination of position hall effect sensors. A sensorless control algorithm has been developed using the zero crossing detection of three line voltages. In comparing with the conventional sensorless algorithms, the proposed method has several advantages including: elimination of the motor neutral voltage, elimination of the fixedphase shift circuit, lower starting speed. The main source of the estimation error of the developed sensorless algorithm is the drop voltage on the stator impedance especially at low speed. However at high speed the estimation error decreases. Therefore, the developed sensorless algorithm is suitable for low-cost application that the load torque is increasing as long as the speed increases.

REFERENCES
[1] P. Pillay, R. Krishnan; #Modeling, simulation, and analysis of permanent-magnet motor drives. II: The Brushless DC Motor Drive$, IEEE Transactions on Industry Applications, vol. 25, No. 2, pp. 274-279, March/April 1989. [2] B.K. Lee, T.H. Kim, M. Ehsani; #On the feasibility of fourswitch three-phase BLDC motor drives for low cost commercial applications: topology and control$, IEEE Transactions on Power Electronics, Vol. 18, No. 1, pp. 164-172, January 2003. [3] J.P. Johnson, M. Ehsani, Y. Guzelgunler; #Review of Sensorless Methods for Brushless DC$, in Proceeding of the 34th IEEE IAS Annual Meeting Conference, 1999, vol.1, pp. 143 & 150. [4] C. Tsung Lin, C. Wen Hung, C. Wen Liu; #Sensorless Control for Four-Switch Three-Phase Brushless DC Motor Drives$, in Proceeding of the Forty-First IEEE Industry Applications Conference Meeting (IAS'06), Taiwan, Oct. 2006, vol. 4, pp. 2049-2053. [5] A. Halvaei Niasar, H. Moghbelli, A. Vahedi; #High Performance Torque Control of Brushless DC Motor Drive based on TMS320LF2407 DSP Controller$, International Review of Electrical Engineering (IREE) Journal, Vol. 2, No. 3, pp. 438447, May-June 2007.

Appendix: BLDC Motor Parameters Pn Tn In R Ls Kt 425 [W] 12 [N.m] 10 [A] 0.64 ['] 1.0 [mH] 1.194 [N.m/A] Zp
n

VDC-bus J M Ke

16 [pole] 350 [rpm] 60 [V] 5e-4 [Kg. m2] 0.25 [mH] 0.0667 [V/rpm]

667