Low Power Extended Range DC Motor Controller
By
Jeffrey Ma
Sungjoon Cho
ECE 345 Senior Design Project
Spring 2004
TA: Joseph Mossoba
May 4, 2004
Project #47
ABSTRACT
The goal of this project is to design and implement a low power extended range DC motor speed controller. The utilization of an electric DC motor with an effective speed control is very extensive in today’s market and is used to run anything from an electric bicycle to a golf cart. A buck-boost converter was designed for an input voltage range between 9-15V, and an output range of 0-15V. The implementation of a speed control allowed the user to control the speed of the motor and to provide speed regulation. The DC motor controller was designed for outputs up to 250W. At the same time, the converter worked well for low power, thus making the range of possible applications very broad.
Table of Contents
Page
1. Introduction ……………………………………………………………………………….1
1.1 Specifications ……………………………………………………………………..1
1.2 Block Diagram ……………………………………………………………………1
1.3 Subprojects ……………………………………………………………………….2
2. Design Procedure …………………………………………………………………………3
2.1 Buck/Boost Converter ……………………………………………………………3
2.2 PWM and Gate Drive …………………………………………………………….5
2.3 Motor ……………………………………………………………………………..5
2.4 Speed Sensor ……………………………………………………………………..6
2.5 Speed Control …………………………………………………………………….8
3. Design Verification ……………………………………………………………………...10
3.1 Equipment …………..…………………………………………………………...10
3.2 Converter verification …………………………………………………………...11
3.3 Overcoming high current ………………………………………………………..14
3.4 Sensing verification……………………………………………………………...16
3.5 Speed Control verification ……………………………………………………....16
4. Cost Analysis …………………………………………………………………………....18
4.1 Parts Cost ………………………………………………………………………..18
4.2 Labor Cost …………………………………………………………………….....18
4.3 Total Cost ………………………………………………………………………..18
5. Conclusions ……………………………………………………………………………...19
5.1 Successes ………………………………………………………………………...19
5.2 Uncertainties …………………………………………………………………….19
5.3 Future Development …………………………………………………………...... 19
References ……………………………………………………………………………………….20
3
1. INTRODUCTION
In today’s market, products with a wide range of use serve a great value. The development of a DC motor controller with an extended range of output power will allow it to be utilized in a wide variety of applications.
The DC motor controller proposed in this design will maintain speed control while having the ability to output a very wide range of current and power.
1.1 Specifications
The prominent specification of this project is the voltage conversion. The input range was designed for 9-15V, representing the inconsistencies of a 12V lead acid battery. The required output range was 0-15V, thus warranting a buck-boost converter. In addition, the output power was required to handle 250W continuously. Finally, a control circuit was implemented to control the load motor speed through an outside reference, while keeping the speed constant through feedback.
1.2 Block Diagram
The basic block diagram of the project is shown in Fig. 1.1.
Figure 1.1. Block Diagram for the Extended Range Motor Control Circuit.
The gate drive and converter both take power from a power supply. The gate drive produces a duty cycle through PWM comparison, and feeds it to the gate of the MOSFET in the converter. The converter output is then measured through a current sense resistor, whose reading is sent to the control circuit. Within the control circuit, the converter output is compared to a user-defined control reference through the use of operational amplifiers. The resulting error signal is then sent back to the gate drive, which compares the error signal to the internal reference, in order to adjust the duty ratio, and thus adjust the converter output.
1.3 Subprojects
Initially, the project was divided into two subprojects consisting of the gate drive/converter aspect, and the motor control/feedback aspect. However, throughout the course of the project, it was discovered that laboratory limitations restricted us from implementing a high power motor control circuit, due to the lack of a high power motor. As a result, the project was split into two different subprojects consisting of entirely separate circuits. One part consisted of high power conversion, and the other dealt with low power motor control and feedback. However, the buck-boost converter and gate drive were identical for both circuits, with the exception of the high current interconnections and an inductor rated for higher current in the high power circuit. The two separate circuits are shown below in Fig. 1.2, with the low powered circuit with motor control built on the power board, shown on the left, and the high powered converter built on a vector board, shown on the right.
Fig. 1.2. Circuits for the two separate low power/high power subprojects
2. DESIGN PROCEDURE
2.1 The Buck-Boost Converter
Figure 2.1. Basic circuit design for the buck-boost converter
The primary concern in the design of this converter was the high current levels that would be present with high power loads. The maximum dc current was calculated through the relation,
Imax = Pmax / V (2.1)
Since the maximum power was specified to be 250W, which would be reached at the highest output voltage of 15V, the maximum dc current was calculated to be 16.7A. However, with this calculated current referring to the output current, the maximum possible current within the circuit had to be compensated even more due to the loss that would inevitably be present in the converter. As a result, we doubled the maximum input current to estimate the input current, and thus rated the circuit for 33.4A.
Although this rating was likely a heavy overshoot of the actual input rating, at the levels of current that we estimated, it was agreed upon that a safe overestimation would be much better than the potentially dangerous consequences of underestimating. In addition, for the high power circuit, efficiency was not a high priority, which made overestimating in design make even more sense.
2.1.1 Switch Selection
With the current rating determined, the first step of the design was to determine the switches that would be used. The fundamental concept of dc-dc conversion is the use of switching to change the voltage from the input to the output. The MOSFET acts as the first switch, which is controlled by the duty cycle supplied by the gate drive. When the square wave duty cycle is high, the positive bias at the gate will cause current to freely flow from the drain terminal to the source terminal. As a result, in this situation the MOSFET can be seen as a short circuit. Due to Kirchoff’s laws, when the MOSFET is on, the diode must be off. Similarly, when the duty cycle is low, the MOSFET becomes an open circuit, and the diode basically acts as a short circuit with the exception of a small voltage drop.
The MOSFET used in this project is the IRFP044N, with ratings of VDSS = 55V, ID = 53A and RDS(on) = 0.02Ω. This particular FET was chosen due to the high current rating, as well as for the relatively low RDS(on). In addition, the TO-247 casing was chosen in order to attach large heat-sinks. Although at first glance the 53A rating may seem to be too severe of an overshoot, the rating is somewhat of a marketing trick, as the 53A rating pertains to an operating temperature of 25°C. Since it is nearly impossible to keep the temperature as low as 25°C at such high current levels, the more realistic rating was the current rating at 100°C, which is 37A, and thus much closer to our estimated maximum
The diode chosen is the MBR4045WT schottky rectifier, with ratings of IF(AV) = 40A, VR = 45V, and VF = 0.56V. Like the FET, the TO-247 package was chosen for heat-sink purposes. Despite the high ratings, the forward voltage drop is very low and leakage is moderate, thus making the MBR4045WT a sensible choice.
2.1.2 Inductor Design
With the switches determined, the next step was the design of the storage components. Switching causes the charging and discharging of the storage elements which essentially determine the average output voltage. The inductor acts as current source when the FET is off and the diode is on, and the output capacitor provides the voltage source to the output when the FET is on and the diode is off.
The two considerations in determining the inductor value of a buck-boost converter circuit are the maximum current ripple, and the avoidance of discontinuous mode. Because the current ripple was not a large concern in the design of this convert, the inductor value was designed to avoid discontinuous mode. Avoiding discontinuous mode is important in order to avoid the resulting control and regulation problems. The general equation for the inductor is
V = L*Δi/Δt (2.2)
Rearranging Eq. (2.2) to find the critical inductance required to keep the inductor current above discontinuous mode yields
Lcrit ≥ Vin(D1T)/Imax (2.3)
D1 refers to the duty ratio required to produce the maximum required 15V. D1 can be estimated by the ideal buck-boost relationship given as
Vout/Vin = D1/(1-D1) (2.4)
Thus, theoretically the maximum duty ratio will be needed when the input is 9V and the desired output is 15V. Using Eq. 2.4
15V/9V = D1/(1-D1) (2.5)
and solving D1 is found to be 0.625.
The only other unknown variable is Eq. 2.3 is T, which refers to the switching period. The switching period is given by the relationship
T = 1/f (2.6)
Thus, to determine the switching period, a switching frequency for the circuit had to be designed. In selecting the switching frequency, two major tradeoffs were considered. As implied in Eq. 2.3, a high switching frequency allows for the required inductance to be low. However, with increasing frequency, higher switching losses result. Due to the high current aspect of our converter, high switching losses were deemed to be unacceptable and a conservative frequency of 35kHz was chosen.
Thus, using Eq. 2.6, we solve
T = 1 / 35000Hz (2.7)
and find the switching period to be 28.6µs.
Having determined the worst case duty ratio, the switching period, and the maximum current, the critical inductance was calculated using Eq. 2.3
Lcrit ≥ 15V(0.625)(28.6µs) / 16.7A (2.8)
Lcrit ≥ 16.1uH (2.9)
Thus, the inductor value was designed to be 25uH.
With the inductor value determined, the next step was to design the fabrication of the inductor. Due to the high current aspect of the project, it was critical to minimize the number of turns of the inductor in order to minimize the wire loss. As a result, the Micrometals model T300-26D toroidal core was chosen for its high inductance/turn2 ratio of 160nH/turn2. Using this core, the number of turns required to ensure a value of 25uH was found using the equation
N = (L/AL)1/2 (2.10)
with AL representing the inductance/turn2 ratio. Thus, the required number of turns was calculated to be 13 turns.
Due to the high levels of current passing through the inductor, along with the relatively long length that the wires would be, it was important to make the windings of the inductor to be rated properly. The thickest wires that were available in the lab at the time of inductor design were 14-gauge wires. However, referring to the AWG table, it was determined that 6-gauge wire would be needed to safely handle the potential current flowing through the inductor [1]. As a result, we decided to braid five 14-gauge wires in parallel to create a 6-gauge wire. This was made possible because each wire in parallel decreases the gauge rating by two [2].
For the low power circuit, an inductor was created with the same core and the same number of turns, but only one 14-gauge wire was used as the windings because of the significantly lower current ratings.
The inductance of the high-power inductor was measured to be 30.8uH, while the low-power inductor had a value of 28.7uH.
2.1.3 Capacitor Design
The primary purpose of the input and output capacitors is to reduce the voltage ripple caused by switching. Equation 2.11 below gives the capacitance required to keep the voltage ripple below a certain value [1].
C = Vout(1-D1)(T)2 / (8*L*ΔV) (2.11)
Using our previously determined inductor value and a reasonable voltage ripple of 2.5%, we calculated our capacitor value to be 49.7µF. Thus, we conservatively chose 100µF electrolytic capacitors for both the input and the output. However, as it will be discussed in the later sections, the chosen capacitors had many problems dealing with the heat, and therefore had to be replaced with larger capacitors.
2.2 The Gate Drive
Initial design and testing was done using a gate drive run by the UC3843 PWM chip. However, with implementation of the control circuit, it was discovered that the UC3843 had an internal current source that pointed in the outward direction of the PWM Feedback input. As a result, the output signal of the control circuit feeding the feedback input was partially negated. Thus, the control circuit had no effect on the duty cycle, and an alternative was needed.
The alternative gate drive chip that was discovered was the TL494 PWM chip. The datasheet for the TL494 also revealed a current source on the Feedback input node, but unlike the UC3843, the current source was pointing inward. The TL494 gate drive chip and external circuit schematic is shown below in Fig. 2.2.
Fig. 2.2. Schematic of the gate drive circuit
The duty cycle was initially designed for manual user control using the internal reference voltage with a potentiometer. However, with the implementation of the control circuit, the control circuit output was applied directly the Feedback PWM Comparator Input at pin 3. The output was made usable for the gate of the MOSFET by including a pull-up scheme using VCC and two 250Ω resistors in parallel. Finally, the frequency was designed to be set by a potentiometer at pin 6.
2.3 Motor
We first needed to find a motor in order to involve the variable speed aspect of our project and to ensure that our control circuit was able to function properly. Although our original converter circuit was designed for a maximum output current of about 20 amps, we were unable to obtain a motor with a current rating that could handle that high of a current without spending hundreds of dollars. As a result, we elected not to test the speed control for high power, but rather focus on developing a functional control circuit. At the same time, we would build a converter that is capable of outputting our desired current levels and implement the converter with the control circuit only for low power.