Low Carbon Footprint Electric Lawn Mower [PDF]

Low Carbon Footprint Electric Lawn Mower Final Report Kraig Kamp David Sharpe Jamin Williams

Advisors: Dr. Huggins Mr. Gutschlag May 16, 2008

Abstract Environmental air pollution and carbon emissions are becoming significant problems. A contributor to this pollution is the use of gasoline-powered lawn mowers. Our overall project goal is to design a battery powered lawn mower and photovoltaic charging system that will diminish emissions. The project consists of two separate systems: a battery-powered lawn mower and a photovoltaic system to charge the battery. The mower will use a microcontroller to control the speed of the cutting blade and display the charge status of the battery. The charger will use the UC3909 battery charging chip to control the charging algorithm for the battery. The system will be designed to be competitive in function and cost versus benefits with gasoline powered lawn mowers.

2

Table of Contents I. Introduction II. Functional Description and Block Diagrams

4 4-7

a. Mower Functional Description

4

b. Charger Functional Description

5

III. Functional Specifications and System Requirements

8-10

a. Mower Specifications

8

b. Charger Specifications

10

IV. Design and Analysis

11-29

a. Battery

11

b. DC Motor Modeling

12

c. Snubber Circuit

15

d. Heat Sink

16

e. PSPICE Simulation

18

f. Mower System Circuit

20

g. Solar Panel Calculations

21

h. Charging Algorithm

22

i. Buck Converter

24

j. UC3909 Equations

26

k. Measuring State of Charge

28

l. Measuring Battery Voltage

28

m. Determining Throttle setting

29

V. Implementation and Results VI. Recommendations for Future Work VII. Applicable Standards and Patents

30 31 33

VIII. Equipment List

36

IX. Bill of Materials

36

X. References

37

Appendix A: UC3909 Design Equations

38-44

Appendix B: Charger Circuit Components

45-47

Appendix C: Software Flowcharts

48-55

3

I. Introduction This report describes the design, implementation and testing of the Low Carbon Footprint Electric Lawn Mower. A top down design approach is followed with functionality first described at the system level including detailed block diagram is developed. Next the design equations are discussed to illustrate more specifically each part of the system, followed by simulations, and final system testing. The report concludes with a parts list, references and final conclusions. II. Functional Description and Block Diagrams The low carbon footprint electric lawn mower consists of two separate systems: a mower block and a charger block, which are shown in figs. 1 and 2. The lawn mower is powered using batteries, and the charger uses a photovoltaic array to recharge the battery. A. Mower Functional Description The main components of the mower, as seen in Fig. 1, are a dc motor, a battery, a controller, and a user interface. The user interface consists of a power switch, a throttle to control the speed of the mower blade, and a safety switch to start and stop the motor. The user interface also consists of a display. The lawn mower uses a 24V DC motor that is powered by two 12V lead-acid batteries. The motor shaft speed is controlled by a microcontroller, which accepts inputs from the user. A PWM signal from the microcontroller is applied to the gate driver which then drives power MOSFETS and controls the average voltage applied to the motor. Detail specifications for the motor and batteries are given in Section III. Figure 2 shows the overall software flowchart for the lawn mower system. Additional flowcharts for each software module are listed in Appendix C. The main purpose of the software is to control motor speed and to turn the motor on or off. To control the motor speed, the software converts a voltage measurement from a potentiometer into a PWM signal. This signal is sent to the motor driver circuitry to change the speed of the motor. The user must engage the safety switch in order to turn the motor on. If the switch is disengaged, the software turns the motor off. The software also handles other tasks. It determines the state-of-charge (SOC) of the battery by utilizing a method known as current counting and uses the SOC to protect the battery from being over-discharged. The software prevents the system from drawing too much current by measuring current via a current shunt. An LCD display is controlled by the software as well. The display provides information about the system to the user such as SOC and terminal voltage of the battery, the throttle setting, and the current that is flowing through the motor.

4

B. Charger Functional Description The charger system, shown in fig. 3, consists of a solar panel which supplies electric power to the charger circuitry which then charges the batteries. The system is designed so that the solar charger can fully charge two completely discharged batteries in at least 5 days. The charger circuit is controlled by the UC3909 charge controller chip. This chip monitors battery voltage and charge current. It varies both voltage and current according to the state of charge of the battery based on the standard charge algorithm sequence for lead acid batteries. This allows for trickle charge, bulk charge, over charge, and float charge as needed to maximize battery life.

Fig. 1 – Mower Block Diagram

5

Fig.2 Overall Software Flowchart

6

Sunlight

Input Voltage Regulation

Solar Energy

Solar Energy

Solar Panel

LM7815 15 VDC

Charger Controller UC3909

Gate Driver 5V PWM

TC4424

Buck Converter 15V PWM

IRF640

DC V/I (Higher Current)

Batteries Voltage/Current Feedback

DCM0035 Charged in Parallel

Fig.3 Solar Charger Block Diagram

7

III. Functional Specifications and Requirements Section II presented a qualitative discussion of general functionality of the mower and solar charger systems. In this section, the various subsystems are described in more detailed including quantitative specifications and functional requirements. A. Mower Specifications The lawnmower is a push-type mower with an electric motor rotating an 18 inch blade to remove 1 ½ to 2 in. off the height of average density grass at a walking speed of approximately 2.66 ft/s. The mower weighs no more than 90 lbs. The motor is powered by batteries with enough capacity to mow a 10,000 sq. ft lawn in one hour. The mower has a power button to power up the controls on the handle and a separate start button to start the electric motor. The circuitry includes over current protection along with a safety switch that must be held down in order to keep the motor running. The batteries that power the mower can be removed from the mower deck but is not necessary while being recharged. Battery: The battery power needed for our application is provided by two 12 volt batteries connected in a series to make 24 volts. The capacity of each battery is 35 AmpHours (AH) so that the mower has enough power to mow a 10,000 sq. ft. yard in one hour. The chemistry make-up of the battery is deep discharge sealed lead acid with a combined weight of approximately 50 lbs. DC Voltage Regulators: The DC voltage regulators convert the variable battery voltage (12-24V) to either 5V or 15V. The 5V regulator is used to power the microprocessor and display and the 15V voltage regulator is used with the gate driver. Controller: The controller is used to start and stop the motor, control the speed of the motor, and control the display. The controller utilizes open-loop methods with a throttle control so the user can set the speed of the motor shaft. The signal that is output to the MOSFET’s is a PWM signal with a frequency of 4 kHz. The controller also monitors the current draw of the motor for over-current protection. The maximum continuous current is 40A for 5 seconds. User Inputs: The user inputs consist of a power button to turn the controller on/off, a start button to start the mower. This button also acts as a safety switch and must be held down in order to keep the blade rotating. A throttle control to vary the speed of the mower blade is also an input.

8

Display: The Optrex 24X2 LCD is the operating display, which is shown in Fig. 4. The display has a battery symbol indicating the charge left on the battery and a speed bar graphic to show the user the relative RPM at which the blade is rotating. In addition, the display also shows the terminal voltage of the batteries and the current flowing through the motor.

Fig. 4 LCD Operating Display Power MOSFETS: The MOSFETS used to drive the motor are two HEXFET IRF044N power MOSFETs wired together in a parallel configuration. The gate of the MOSFET receives a PWM input signal at a frequency of 4 kHz. The MOSFETs are protected with a snubber circuit that keeps voltages under the Vds rating of 55V. Heat Sinks: Two Wakefield 657-15AEPN heat sinks are attached to the MOSFETs and one Wakefield 287-1ABE heat sink is attached to the free wheeling diode in the snubber circuit. These heat sinks maintain safe thermal operating temperatures for these components. Gate driver chip: The TC4424 takes a 0-5 Volt signal from the microcontroller and outputs a 0-15 Volt PWM signal to drive the gate of the MOSFET at a switching rate of 4 kHz. Motor: The motor rotating the blade is the Tecumseh 90000A permanent magnet reversible motor with an input voltage of 24 VDC and generates 1.54 HP at 3200 RPM. These specifications provide the power to rotate an 18 inch blade at a sufficient velocity to cut 1 ½ to 2 inches off the top of average density grass. Over current Protection: A 40 amp circuit breaker is used for over current protection on the battery for testing and a 30 amp inline fuse is placed between the positive terminal of the battery and the positive terminal of the Tecumseh 90000A DC motor. Disconnect Switch: A disconnect switch is essential in the circuitry. In the case that a MOSFET fails while mowing and acts as a short circuit, the motor still runs and cannot be turned off by normal procedures. This manual disconnect switch is mounted by the handle and breaks the circuit so the blade stops spinning.

9

B. Charger Specifications The charger uses energy collected by a solar panel and directly transfers the energy to the battery. The system charges a maximum of two 12 Volt batteries at one time, wired in a parallel configuration. The charger needs 5 days at 530 kJ/day of solar energy to charge the two batteries to full capacity. The charger utilizes the Texas Instruments UC3909 lead-acid battery charger chip to control the charging process. 15V Voltage Regulator: The DC regulator for charger electronics regulates the DC power input from the solar panel. This regulator powers the UC3909 and the buck converter gate drive chip Gate driver chip: The Charger system utilizes the same gate drive chip as mentioned in the Mower specifications because it is a dual input chip. The gate driver takes the PWM input from the UC3909 and converts it to 15V for the gate of the buck converter MOSFET. Solar Power: The solar power is collected with a BP350 50W panel with an open-circuit voltage of 21.8 V and a short circuit current of 3.17 A. The panel can collect at least 530 kJ of energy per day to charge the two fully discharge 12V batteries to maximum capacity in 5 days. Charger Chip: The Solar Charger System is controlled by the UC3909 Switch mode lead acid battery charger chip. This chip is used in conjunction with the buck converter in the charger circuit. This method and the charge algorithm is further discussed in section IV, part G. Buck Converter: Because the charger controller chip uses a PWM signal to vary voltage and current applied to the batteries, a buck converter is necessary to ensure that voltage and current is constant even though the input is switching. This is further discussed in Section IV, part H. Batteries: The battery block consists of two 12V, 35 AH batteries, as discussed in the lawnmower subsystem breakdown but is wired in parallel for charging.

10

IV. Design and Analysis Section III talked about the detailed subsystems of the mower and charger. This next section describes the design and analysis of several subsystems and how they were implemented into the project. A. Battery To meet the specification of mowing a 10,000 square foot yard in 1 hour, the batteries must be able to power the motor at full speed for the entire time. To size the batteries appropriately, tests were carried out. This testing revealed that the average running current of the mower system was approximately 18A. According to Fig. 5, at a discharge rate of about 20A, a 35Ahr battery will last about an hour. After that, the terminal voltage drops off abruptly, indicating a fully discharged battery. It is important to note that because the mower uses a 24V DC motor, it is necessary to use two 12V batteries in series, to obtain this voltage. Because the batteries are in series, the current through them will be the same, thus both batteries must have the 35Ah (18Ah @ 20A rate) capacity rating to provide enough energy.

Fig. 5 – 12V, 35Ahr Lead Acid Battery Discharge Characteristics

11

B. DC Motor Modeling In order to draw a schematic of the power electronics and the Tecumseh 90000A in PSPICE, certain motor parameters must be determined. These parameters are the resistance in the windings, the motor torque constant, the viscous and static friction coefficients, and the mass moment of inertia. These mechanical parameters can be simulated as electrical circuits in PSPICE as shown in Fig. 8. The mass moment of inertia is modeled as an inductance in the model and the friction coefficients are modeled as a resistance. Torque can be modeled as a currentdependent voltage source. It is proportional to the current flowing through the motor. The following are steps taken to measure these parameters: 1. Measure Ra which represents the resistance in the windings To do this, a voltage ( Vs ) is applied that is low enough such that the shaft of the motor does not spin, so as not to introduce the back EMF Voltage created by the motor. A current probe is used to measure the current and then Ra is determined from Ohm’s law.

Fig. 6 Schematic used to find Ra

Ra =

Vs Ia

Eqn. 1

2. Find the motor constant K E = K t The input voltage in Fig. 6 is set to ½ of the rate input voltage of the motor ( Vs =12 V) and I a is measured, then the speed of the motor shaft in radians/second is also measured.

Ke = Kt =

V s − I a Ra

ωs

Eqn. 2

12

3. Find the static friction coefficient, TS .F . and the viscous friction coefficient b.

Vs is first set to 8 volts and then 16 volts and ω s and I a are measured at each voltage. Then, the following equation is used to find both unknowns.

Tdeveloped − TS . F . − bω s = K T I a − TS .F . − bωs = 0

Eqn. 3

4. Find the mass moment of inertia, J. This is accomplished with a coast down test by driving the motor at 24 volts. Once the shaft is up to speed the power is shut off and the motor voltage decay is captured with the oscilloscope: 25

20

15

10

5

0

-5 -1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Fig. 7 Motor Coast curve averaged in MATLAB

Voltage is on the y-axis and time is represented on the x-axis. Using this plot‘t’ is found ,which is the time the motor took to coast down from 24 volts to 0 volts. By the graph above t=.847 seconds. Using the next equations, the final parameter, J is determined.

 V TS . F .  −τt ω (t ) =  + e K b  t 

Eqn. 4a

13

Setting ω (t ) = 0 , τ can be found which will be used to find J. J = b * τ

Eqn. 4b

(Note: this can be checked at any speed, find V at any speed and set ω (t ) = to that particular speed and the right side of the equation should approximately equal ω (t ). )

Using the four steps above the following values were measured and used to implement in the PSICE model shown in Fig. 8. Ra = RM = .0825Ω N *m A = 0.272504 N * m

K t = K e = 0.068723 TS . F .

N *m rad / sec J = LJ = 0.000912kg * m

b = RB = 0.000535 LM = 153µH

Here, LM is inductance of the motor terminals measured with an LRC meter.

Fig. 8 PSPICE DC Motor Model (www.ecircuitcenter.com)

14

C. Snubber Circuit As stated earlier, a snubber circuit is required to protect the MOSFETs from being damaged when large currents could flow into the drain. This is a problem because the maximum voltage from the drain to source cannot exceed 55V for the MOSFETs. Diode 1 on the left side of the circuit turns on and lets the current flow through it when the MOSFET switches off. The circuitry on the right side of the circuit essentially gives Diode1 time to turn on all the way and accommodate the high currents. Below is a schematic of the snubber circuit To negative terminal of the motor

Rs To positve terminal of the motor D1

D2

Cs

To Drain on the MOSFET

To Source on the MOSFET

Fig 9 Snubber Circuit

In order to obtain values for Rs and C s the following constraints need to be taken into consideration. First, Rs should be small enough so that the largest current that could be flowing through the motor multiplied by Rs will not exceed the Vds rating of 55 volts. Therefore, approximately 400 amps is the greatest amount of current flowing though the motor according to simulations so, V=I*R 55V = 400*R

Eqn. 5

R=.135 ohms.

15

In addition, Cs is determined using Eqn. 6 as follows.

Cs =

I Do * t f 12 * Voff

Eqn. 6

Where I Do is the max current of 400 amps, t f is the switching frequency of the MOSFETs (4 kHz), and Voff = Vs + V D1on or 24 V+.4512 V=24.4512V.

Eqn. 7

(Note: V D1on is the voltage drop across the diode) D. Heat Sink While making initial tests in the motor driver circuit it was noticed that the IRFP044N power MOSFETS were becoming extremely hot. Looking into the problem these calculations were made to find out the junction temperature of the MOSFETS. Using the IRFP044N data sheets the following data was obtained.

VDSS = 55V RDS ( on ) = 0.020Ω I D = 55 A °C W °C RϑCS = 0.24 W °C RθJA = 40 W RθJC = 1.3

Also, datasheet supplies a Normalized On-Resistance Chart at a junction temperature of 100 °C which yields R DS (on ) = 1.5Ω . This is normalized so this value needs to be multiplied by the R DS (on ) of 0.020 Ω to get .03 Ω . Next, the power dissipation needs to be calculated, and the average drain current must be known. The mower system will have a worst case scenario of 40 amps flowing through the drain at normal operating procedures and since there are 2 MOSFETS in parallel the current will be evenly distributed through the both of them. So the drain current is 20 Amps, therefore:

PD = I D2 * R DS ( on ) = 20 2 * 0.3 = 12W

Eqn. 8

16

Now the junction temperature is: TJ = PD RθJA + T A = 12 * 40 + 37.7 = 517.7°C

Eqn. 9

Note: T A was chosen by the design team to be 37.7 °C . 517.7 °C is not an acceptable temperature for the IRFP044N power MOSFET, so a heat sink is needed. Using a heat sink: TJ = PD RθT + T A where RθT = RθJC + RθCS + RθSA

Eqn. 10

RθSA can be found by using the natural convection characteristic chart of a specified heat sink found on the data sheets. The heat sink chosen for this application is the Wakefield 657-15ABEP and at a power dissipation of 12 W the heat sink temperature rise above ambient is about 58 °C which yields: RθSA =

∆TSA 58°C °C = = 4.83 PD W 12W

RθT = 1.3 + 0.24 + 4.83 = 6.37

°C W

Eqn. 11

Eqn. 12

So now the new junction temperature can be calculated

TJ = 12 * 6.37 + 37.7 = 114.14°C .

Eqn. 13

114.14 °C is within the 175 °C maximum operating junction temperature of the IRFP044N power MOSFET.

17

E. PSPICE Simulation Having determined the electrical and mechanical properties of the motor as well as the components of the snubber circuit, a mower simulation circuit can be developed. The following circuit diagram is drawn in PSPICE to simulate the basic operating conditions of the mower motor control system.

Ra .0825 2 Motor 153uH

Motor Electrical Model

1

Motor Mechanical Model

EMF + -

+

H1

D4

R1

MUR405

.14 I

D3

V2

V+

Drain

24Vdc FET1

FET2

MUR405

I

IRFP044N V1 = 0 V2 = 15 TD = 0 TR = 10n TF = 10n PW = .0005 PER = .001

PWM

C1 8u

IRFP044N Source

2 Torque

LJ I 912uH

+ -

H2

V-

1 RB .000535

OPEN 50Meg

OPEN2 50Meg

0

Fig. 10 Motor Model, Snubber Circuit, and MOSFETs Simulation Circuit

18

Fig. 11 Simulation Results of Circuit in Fig. 10 The above simulation done in PSPICE shows the voltage and current curves with respect to the motor shaft speed which is the curve in blue. As seen above the teal curve represents the current flowing through the motor. At initial start up the inrush current is around 400 amps and as the shaft approaches top speed the current decreases to about 3 amps. The green curve shows the current through the freewheeling diode 1 in the snubber circuit. In this particular simulation the current is measured through one of the MOSFETS instead of measuring it through the two of them so the simulated current is double of the motor current. The magenta curve shows the voltage with respect to speed and as shown it stays at a constant 24V. Note that the simulation does not take into consideration all of the motor losses, so the simulation shows the motor pulling less current, but the simulation shaft speed was almost the same as the experimental.

19

F. Mower System Circuit After simulations and software programming the mower system circuit was implemented. The system schematic is shown in Fig. 12 below.

12V

12V

7815

24V Batteries

I G O 7805

.33u

M

I G O 220u/50V .33u

1kΩ 1kΩ STPS20120D

Safety Switch

217Ω 10kΩ

2 3 5 7 8 9 10

Vcont

VCC

R/W D0

E

LCD

D1

D7 D6

D2 D3

RS

D5 GND 1

D4

4

1

6

2

14

3

13

4

12

5

11

6 7 8 9 10 11 12 13 14

PC6 PD0

PC5

Atmega 168

PC4

PD1

PC3

PD2

PC2

PD3 PD4 VCC

PC1 PC0 GND

GND

AREF

PB6

AVCC

PB7

PB5

PD5

PB4

PD6

PB3

PD7

PB2

PB0

PB1

28 27 26

1

25

2

24

3

23

4

22

NC

NC

In A

Out A

GND

VCC

In B

Out B

5 6 7 8

21 20 19 18 17 16

.1Ω

15

IRFP044N IRFP044N From UC3909 Stat 0 From UC3909 Stat 1

STPS20120D .68u

.68u

.005Ω 5W

Fig. 12 Mower System Circuit Diagram

The mower circuit above contains all of the circuitry used to power and control the mower system. The controller (Atmega 168) software flowcharts are further discussed in Appendix C. The circuit also references the UC3909, which is discussed in part I of this section. 20

G. Solar Panel Calculations The charger system is powered by a photovoltaic panel. This component is the most expensive part of the entire project, so the smallest size that can charge the batteries in 5 days must be used. The solar panel must be able to supply enough voltage and current to charge the two 12V batteries, as well as collect enough energy in 5 days to do this. A 50W solar panel, the BP350, was the most appropriate solar panel to meet the requirements. It is important to note that in the solar charger system, the batteries are charged in parallel, so that they can be charged at 12V, because to charge at 24V, it would require a much larger solar panel. The solar panel is selected based on the minimum solar energy the U.S receives in places that can still grow grass. According to Fig. 13 from the National Renewable Energy Laboratory website, the upper parts of the country receive the least amount of solar energy, and therefore this number should be used when calculating the amount of solar energy the panel can collect.

Fig. 13 Minimum Daily Solar Radiation Per Month (NREL.gov)

The efficiency of the BP350 is 10%, which means 10% of the solar radiation energy incident on the panel collected is converted to electric energy. Taking into account this efficiency combined with the panel area and minimum radiation of 3.5 KWh/m2/day yields the energy collected per day by the panel.

21

 .092903m 2  3.5KWh  1000W  1J / s  3600s   2  4 ft 2    (10% ) = 0.4682MJ / day Eqn. 14 2 1 / 1 1 ft m day KW W hr          

Another important note is that the solar energy the US receives varies monthly based on the seasons. So given all the above information, a charge time per month chart was developed. Table 1 Monthly Battery Charge Times 2

Actual days to charge 2 - 35AH batteries

KW-Hrs/ m /day Solar Energy Emitted

KJ / day of Solar Energy Collected

January

2.0

353

8.6

February

3.0

530

5.7

March

4.0

706

4.3

April

4.0

706

4.3

May

5.0

883

3.4

June

5.0

883

3.4

July

5.0

883

3.4

August

5.0

883

3.4

September

4.0

706

4.3

October

3.0

530

5.7

November

2.0

353

8.6

December

1.0

177

17.1

Month

From the data presented in table 1, it is evident that the 50W solar panel works for this project. During the grass cutting months of March through September, the batteries can be charged in under 5 days. The winter months have a much higher charge time because less the US receives less solar energy. This is acceptable because the charger system only needs to keep a float charge on the batteries while they are not used during the winter. H. Charging Algorithm Now that the solar panel has been selected, it is interfaced with the charger circuit. But first, some of the basics of charging a lead acid battery must be discussed. Lead acid batteries are charged in different stages because of the nature of the battery chemistry. The charging stages can be seen on the following figure taken from the technical paper by Laszlo Balogh.

22

| Trickle |

Bulk

| Over |

Float

|

Fig. 14 Charge Algorithm Voltage and Current Characteristics

1.

Trickle Charge

This is the first stage of charging, where the battery is completely discharged. Here, battery current kept low and constant in order to bring the terminal voltage high enough to start the next stage of charging. 2.

Bulk Charge

This is the high current stage where most of the battery’s charge is returned. Here battery current is kept constant while terminal voltage gradually increases further until it reaches the over charge cut off voltage.

3.

Over-Charge

This stage is constant voltage stage where the remaining charge is returned to the battery. The voltage is held a couple volts higher than the rated operating voltage, but 23

only for a short period of time. Again, the charge algorithm is based on the internal chemistry characteristics of lead acid batteries. 4.

Float Charge

This is the final stage of the battery charge algorithm. Once the charger reaches this stage, the battery is fully charged. This is a constant voltage stage just slightly over the operating terminal voltage. This stage just maintains the battery so it is ready for use by charging at the same rate the battery naturally self-discharges. I. Buck Converter The charger circuit is controlled by the UC3909 Switchmode lead acid battery charger chip. The circuit consists of the controller, voltage dividers for voltage and current, input voltage regulation, and a buck converter to supply constant current and voltage to the batteries. The controller controls the current and voltage via PWM signal sent to the buck converter’s MOSFET. The basic concept of a buck convert can be seen in the following circuit (wikipedia.org):

Fig. 15 Basic Buck Converter Circuit

The red parts of the circuit represent current flow. The switching is done by a MOSFET that is connected directly to the solar panel. When the switch is on, the source both supplies power to the load as well as the inductor and capacitor. When the switch is 24

off, the diode conducts to complete the circuit, and the energy stored in the inductor and capacitor is then supplied to the source. This effectively provides constant current and voltage even though the source is switching as seen in the next figure (wikipedia.org).

Fig. 16 Buck Converter Characteristics The figure above illustrates the constant current and voltage. The UC3909 can control the current and voltage by varying the duty cycle of the switching waveform. A higher duty cycle means the MOSFET is on longer, thus the average current and voltage are higher, and the converse is true for a lower duty cycle.

25

J. UC3909 Circuit Equations Now that the basics of the buck converter operation have been discussed, it is necessary show how the UC3909 charger circuit works. The controller measures the battery terminal voltage as well as the current through a current sense resistor. The controller also must know what voltage and current cutoff points are in order to change to the next stage of the charging algorithm previously mentioned. It does this by various resistor and capacitor networks. The formulas for calculating these values and the rest of the components in the charger circuit are provided in a technical paper published by the chip’s manufacturer (Balogh). See the UC3909 Battery Charger Appendix A for the formulas required to design the charger circuit. All the charger circuit equations are based on the parameters inherent to the battery. Again, it is important to note that because the batteries are charged in parallel, the voltage is the same at 12V, but the capacity must be doubled to obtain the correct charging currents. Using the characteristics of the batteries selected for this project, the DCM0035 by Interstate, the component values were calculated using the extensive design equations given in Appendix A. The components are connected as shown to the UC3909 as shown in Fig. 17, which is the complete charger circuit.

26

Fig. 17 UC3909 Charger Circuit

27

K. Measuring State of Charge (SOC) As discussed in the Functional Description section, the microcontroller on the mower displays the state of charge of the battery. This requires that the State of Charge (SOC) be measured. Of all the methods researched, current counting (Zhu), as given in Eqn. 15, is implemented on mower controller to determine the SOC.   1 SOC = SOC0 ±  ∗ ∫ Idt    Capacity

Eqn. 15

SOC is the calculated state-of-charge of the battery, SOC0 is the initial state-of-charge of the battery, Capacity is given by the battery manufacturer, and current (I) is measured. Current is measured using a current sense resistor. These are typically resistors that are very small and accurate. A current sense resistor is placed in the ground path of the circuit and the voltage drop was measured across it. Using a .005Ω resistor made it possible to omit a conversion step to determine the actual current because the A/D register was very close to the actual current. Vsense = I ∗ Rsense Imeasured =

Vsense ∗1024 V AREF

Eqn. 16 Eqn. 17

Ex. I=30A Rsense=.005Ω VAREF=5Vdc Vsense=30*.005=.15 Imeasured=30.72

Eqn. 18

L. Measuring Battery Voltage The microcontroller on the mower also displays the terminal voltage of the batteries which requires measurement of the voltage. The maximum voltage of each battery at full charge is approximately 14Vdc. Since the batteries are in a series configuration, the maximum total battery voltage is 28Vdc. Since the A/D channels on the microcontroller can only measure 0-5Vdc, a voltage divider circuit must be used. The maximum output voltage of the divider circuit needs to be 5Vdc. This corresponds to 28Vdc as well as the maximum value that the A/D register can store (210-1=1023 = 5V when VAREF=5V).

28

R2 5Vdc = 28Vdc R1 + R2

Eqn. 19

A 1kΩ was selected for R1 to limit current at the microcontroller. This yields a value of 217Ω for R2.

V batt(Register Value) =

Vbatt ∗ 217 210 ∗ 1000 + 217 V AREF

Eqn. 20

If Vbatt = 28Vdc and VAREF = 5Vdc, then V batt(Register Value) is1022. V batt(Register Value) is divided by the actual voltage to determine how they are related.

1022 = 36.5 28

Eqn. 21

Eqn. 20 is used to convert V batt(Register Value) to the actual voltage on the microcontroller. The microcontroller can only divide by whole numbers. To fix this, it is multiplied by 10 then divided by 10*36.5=365.

V

batt(Register Value)

365

∗10

= V batt(actual).

Eqn. 22

M. Determining the Throttle Setting A voltage divider with a 10kΩ potentiometer is used as the user input for the throttle. The input voltage to the divider is 5Vdc. The output voltage will be between 0 and 5Vdc. 0Vdc will correspond to a 50% speed setting, and 5Vdc will correspond to a 100% speed setting. Timer1 on the microcontroller controls the PWM signal. A counter is incremented until it equals 2046 and then resets to 0. Every time the counter is incremented, the count is compared to a compare register. If they are equal, a pin is toggled. This creates the PWM signal. For example, if the compare register had 1023 stored, this would result in a 50% duty cycle. In order to have 0Vdc correspond to a 50% duty cycle, 1023 should be added to the A/D register value and then stored in the compare register. Compare Register =

Vout ∗1024 + 1023 V AREF

Eqn. 23

29

V.

Implementation and Results

Mower System Results The various subsystems of the mower system and charger system, discussed in Section IV, were implemented and tested. In case of the motor control subsystem, the microcontroller PWM signal with throttle control input, voltage regulators, gate driver and power MOSFETs were successfully implemented. The motor control subsystem was first tested in the lab using a PWM signal supplied by a function generator and system operation was verified. Next the microcontroller was interfaced to the gate driver and speed control of the motor using the throttle input to the microcontroller was verified. With these successful tests, the complete motor control subsystem was finally mounted on the mower platform with the Tecumseh motor with blade and tested by mowing grass. The test was carried out by mowing dense wet grass and cutting 1 ½ inches off the top of the grass. The pace at which the mower was pushed was at a walking speed of approximately 2.66 ft/s. The mower ran for almost an hour and a half and mowed approximately 13,000 sq. ft, before the batteries were completely discharged. This mowing test exceeded the requirement to mow a 10,000 sq. ft yard in one hour. As far as the weight expectation, the final mower system weighed 91 pounds as a prototype. The 90 pound specification certainly could be met by improved mower deck design, optimizing battery size and weight and using a brushless DC motor. Furthermore, the speed control worked correctly. When the throttle knob was adjusted, the motor RPM would decrease or increase as the microcontroller varied the PWM signal accordingly. This change in speed was also displayed correctly. Unfortunately the current and state of charge were not measured or displayed correctly due to problems with the current sense resistor. Though the SOC algorithm was correct, the current measurement was wrong precluding a correct SOC calculation. The problem with the current measurement was noise due to the small current sense resistor. The problem with the voltage display is due to the battery voltage divider not working correctly. Solar Charger System Results The solar charger system was implemented as shown in figure 17. The solar panel input was simulated by the Agilent power supply, and the battery was simulated as a load resister. First, the UC3909 chip operation was verified by supplying an input voltage similar to the solar panel output at peak power, about 19V. Then the PWM signal was measured with the oscilloscope along with the timing capacitor for the switching frequency. Once they were verified, the output of the UC3909 was interfaced to the buck converter which was loaded by the resistor. Then, voltage and current measurements were made on the load. The input voltage was slowly increased until peak power to simulate how the sun moves from dawn to the peak power point. The Chip turned on around 8V, and the load voltage increased to about 10.5V, and current was about 150 mA when the input voltage was at 19V. Once the load voltage reached this level, the charger system regulated it there for further increases of the input voltage up to the power supply limit of 26V. This indicates the charger system was functioning, but not within the specifications necessary to charge two 12V batteries. This is most likely due to the

30

tolerance of the components calculated in appendices A and B. These components are used to set the voltage and current levels for the charge algorithm. Another issue may be noise. Because the UC3909 is analog, it is more susceptible to noise issues, and most of the measurements made were fairly noisy. Similar to the current sense issue mentioned for the mower system, the charger system also uses a low value resistor for current sensing. VI. Recommendations for Future Work

In conclusion, the mower exceeded most of the initial specifications and the charger system remained out of the initial specifications. This project is interesting and offers the opportunity for future work as follows. Mower System Recommendations •

Design the mower with a brushless DC motor. This would dramatically increase efficiency and therefore battery sizing and other parts of the power electronics may have to be redesigned

•

Add a self propelling functionality, because of the increased weight of the batteries. This change would also need further design considerations for the battery size and power usage.

•

Utilize a Hall Effect sensor for measuring motor current. This would increase the accuracy of sensing current for the SOC algorithm.

•

Implement the AC power back up. This was included in the original specifications to allow for quick charging and cases of prolonged cloudy weather, but it was omitted to reduce the scope of the project

•

Obtain a sturdier mower deck. This project used a plastic deck that was not designed to work with the motor and batteries in the system, so it was awkward to push around.

31

Charger System Recommendations •

Redesign and simulate the charger circuit. Because the circuit displayed correct functionality in the laboratory, it appears the problem is with the choice of components. It is recommend that new component values computed and simulated. However this will require a model for a lead acid battery.

•

Interface the redesigned charger circuit to the solar panel.

•

Develop a digital charging system. A digital system would be much less susceptible to noise, and it could be implemented on the same microcontroller used in the mower system. This may turn out to be a project in itself because it will most likely need a closed loop feedback controls design to integrate the charge algorithm.

32

VII. Applicable Standards and Patents

Standards Document # ASAE S440.3 MAR2005

Title Safety for Powered Lawn and Garden Equipment

Developer

UL 1447 (Ed. 4)

Standard for Electric Lawn Mowers The Standard for Safety for Electric Lawn Mowers

UL

UL 1447-2006 UL 82 (Ed. 6) IEC 62093 Ed. 1.0 b:2005 IEC 60086-1 Ed. 10.0 b:2007 IEC 60086-2 Ed. 11.0 b:2007 Electricity. Magnetism. General Aspects (IEC) Other Standards Related to Electricity and Magnetism (IEC)

ASABE

UL

Standard for Electric Gardening Appliances UL Balance-of-system components for photovoltaic systems - Design qualification natural environments IEC Primary batteries - Part 1: General IEC Primary batteries - Part 2: Physical and electrical specifications IEC Electricity. Magnetism. General Aspects Collection IEC

Other Standards Related to Electricity and Magnetism Collection IEC Electric cables - Calculation of the current rating IEC 60287-1-1 Ed. 2.0 - Part 1-1: Current rating equations (100 % load b:2006 factor) and calculation of losses - General IEC Electromagnetic compatibility - Requirements for household appliances, electric tools and CISPR 14-2 Ed. 1.1 similar apparatus - Part 2: Immunity - Product b:2001 family standard IEC IEC 60730-2-10 Ed. 2.0 Automatic electrical controls for household and b:2006 similar use - Part 2-10: Particular IEC C 4512

Small Switches for Single-Phase Motors (E) KSA Automatic Electrical Controls for Household and Similar Use; Part 2: Particular Requirements UL 60730-2-10A (Ed. 1) for Motor Starting Relays UL DC ferrite permanent magnet motors (TEXT OF GB/T 6656-1986 DOCUMENT IS IN CHINESE) SPC UL 1004 (Ed. 5)

Standard for Electric Motors

UL

33

UL 1012 (Ed. 7)

Standard for Power Units Other Than Class 2 UL Household and similar electrical appliances IEC 60335-2-29 Ed. 4.1 Safety - Part 2-29: Particular requirements for b:2004 battery chargers IEC Primary batteries - Summary of research and IEC/TR 61955 Ed. 1.0 actions limiting risks to reversed installation of en:1998 primary batteries IEC A-20

Battery Charging Devices

TA-27

Batteries and Battery Chargers ABYC BATTERY CHARGER (FOR 6/12/18/24 VOLTS) US DoD Balance-of-system components for photovoltaic systems - Design qualification natural environments IEC MOWER, LAWN, ROTARY, WALK BEHIND (HAND PROPELLED WITH BLADE STOP) (NO S/S DOCUMENT) US DoD Circuit-breakers for over current protection for household and similar installations - Part 2: Circuit-breakers for a.c. and d.c. operation IEC

A-A-1741 IEC 62093 Ed. 1.0 b:2005

A-A-744 NOT 1 IEC 60931-3 Ed. 1.0 b:1996

ABYC

34

Patents

Patent Number

Description

US Pat. 4987729

Lawn Mower w/ solar panel attached

US Pat. 4942723

Lawn Mower w/ solar panel attached

US Pat. 5906088

Lawn Mower w/ solar panel attached

US Pat. 5084664

Solar Powered Lead-acid battery charger

US Pat. 4871959

Solar Powered Lead-acid trickle charger

US Pat. 6236175

Process and device for detecting the speed of rotation of a DC electric motor

US Pat. 5321627

Battery monitor and method for providing operating parameters

US Pat. 5656920

Method and apparatus for charging a lead-acid battery

35

VIII. Equipment List

Test Equipment: • • • • • • • • • •

Agilent DC Power Supply E3634A Agilent DC Power Supply 3630A GE DC Ammeter Pioneer Digital Photo Tach DT-36M Power Patrol SLA1079 Sealed Lead Acid Batteries GME PM89 Ammeter and Current Shunt Fluke 87 Digital Multimeter Fluke 45 Digital Multimeter Fluke 337 Clamp Ammeter Tektronix TDS3012B Digital Oscilloscope

IX. Bill of Materials Part

Part Number

DC Motor 24V Battery 50W Solar Panel Mower Frame Motor MOSFET Buck MOSFET MOSFET Heat Sink Diode Heat Sink 30A Fuse Fuse Holder Std. 0.25W Resistors Std. Ceramic Capacitors Speed Potentiometer Safety Switch Buck Inductor Input Capacitor Output Capacitor 5V Regulator 15V Regulator Gate Driver Buck Diode Protection Diode Microcontroller Charger Controller LCD

Tecumseh 9000A Interstate DCM0035 BP350 (estimate) IRFP044N IRF640 Wakefield 657-15ABPN Wakefield 287-1ABE Little Fuse 0297030.WXNV Little Fuse 01530009Z (various) (various) 3852A-282-103AL 8125SHZBE 2216-V-RC UHE1H681MHD UHE1E471MPD LM7805 LM7815 TC4424 MUR405 STPS20120 ATMEGA 168 UC3909 MDLS-24269-HT-HV-S

Total Parts Cost

Bulk Unit Cost 57 26 269 20 1.3398 1.799 0.98 0.659 0.2973 0.6627 0.00855 0.027 4.27 2.5558 1.7085 0.2241 0.1232 0.22828 0.252 1.33 0.14204 0.493 2.39 3.721 5

Quantity 1 2 1 1 2 1 3 1 1 1 20 13 1 1 1 1 1 1 1 1 1 5 1 1 1

Line Cost $57.00 $52.00 $269.00 $20.00 $2.68 $1.80 $2.94 $0.66 $0.30 $0.66 $0.17 $0.35 $4.27 $2.56 $1.71 $0.22 $0.12 $0.23 $0.25 $1.33 $0.14 $2.47 $2.39 $3.72 $5.00 $431.97

36

X. References

http://www.ecircuitcenter.com http://www.mindfully.org http://www.batteryuniversity.com http://www.wikipedia.org http://NREL.gov Balogh, Laszlo. “Implementing Multi-State Charge Algorithm with the UC3909 Swichmode Lead Acid Battery Charger Controller.” (Unitrode) Texas Instruments. 1999.

Zhu, C.B.; Coleman, M.; Hurley, W.G. “State of Charge Determination in a Lead-acid battery: combined EMF estimation and Ah-balance approach” Power Electronics Specialists Conference, 2004, PESC 04. 2004 IEEE 35th Annual vol.3,20-25 June 2004 pp.1908 – 1914

37

Appendix A

Battery Data Equations (Unitrode)

38

Appendix A

Buck Converter Operating Parameters (Unitrode)

39

Appendix A

Power Stage Design Equations (Unitrode)

40

Appendix A

41

Appendix A

Controller Design Equations (Unitrode)

42

Appendix A

Controller Design Equations Cont. (Unitrode)

43

Appendix A

Controller Design Equations Cont. (Unitrode)

44

Appendix B

Charger Circuit Components List Parameter

Description

Battery Data for DCM0035 Lead-Acid Battery V Nominal Battery Voltage NC Number of Cells Crate Vc Vc,max Vc,min Itrickle Ibulk IOCT TC Tmin Tmax Vbat Vbat,min Vbat,max Pch,max

Battery Capacity Cell Float Voltage Maximum Cell Voltage Minimum Cell voltage Trickle Charge Current Limit Bulk Charge Current Limit Over-charge Terminate Current Threshold Cell Voltage Temperature Coefficient Minimum Operating Battery Temperature Maximum Operating Battery Temperature Battery Float Voltage Minimum Battery Voltage Maximum Battery Voltage Maximum Output Power

Value/Part#

Unit

12 6

Vdc

3.6 2.25 2.483 1.75 0.036 1.8 0.45 -0.0035 -23 60 13.5 9.765 15.906 28.6308

Ah Vdc Vdc Vdc A A A V/C C C Vdc Vdc Vdc W

12 22 50000 0.59 0.73 1.353181461 0.487681478

Vdc Vdc Hz Vdc Vdc

Buck Converter Operating Parameters Vin,min Vin,max fs Vd1f Vd2f Dmax Dmin

Minimum Input Voltage Maximum Input Voltage Switching Frequency D1 Forward Voltage Drop (estimate) D2 Forward Voltage Drop (estimate) Maximum Duty Ratio Minimum Duty Ratio

Buck Converter Power Stage Components Design Sheet VRMM (D1) IO,MIN (D1) D1 PD1 VRMM (D2) IO,MIN (D2) D2 tRR IRRM PD2

Diode Breakdown Voltage Diode Current Rating Discharge Protection Diode Diode Power Dissipation Diode Breakdown Voltage Diode Current Rating Buck Freewheeling Diode Diode Reverse Recovery Time Diode Peak Reverse Recovery Current Diode Power Dissipation

23.859

V

3.6 MUR405 1.062

A

33

V

3.6 MUR405 3.50E-008 0.5 0.677999038

A

W

s A W

45

VDSS (Q1) ID,MIN (Q1) Q1 RDSON (Q1) COSS (Q1) IGATE QGS (Q1) QGD (Q1) tOFF; tON PQ1 PHS DIL,MAX L1 IL1,PEAK L1 VC3 IC3,RMS C3 C18 VC5 IC5,RMS C5 C5 RC5,ESR PSN,MAX VC4 C4 C4 R3 R3 PR4,MAX R4 R4 F1

Switch Breakdown Voltage Transistor Current Rating Buck Main Switch Switch ON Resistance Drain Source Capacitance Gate Charge/Discharge Gate-To-Source Charge Gate-To-Drain Charge Approximate Switching Times Switch Power Dissipation Heat sink Power Dissipation Inductor Ripple Current Buck Inductance Inductor Peak Current Buck Filter Inductor Input Capacitor Voltage Rating Input Capacitor RMS current Input Capacitor (electrolytic) High Frequency Bypass For Switches Output Capacitor Voltage Rating Output Capacitor RMS Current Output Capacitor (electrolytic) Output Capacitor’s ESR Snubber Power Dissipation Snubber Capacitor Voltage Rating Snubber Capacitor(polypropylene or metalized film) Snubber Capacitor(polypropylene or metalized film) Snubber Resistor (non-inductive) Snubber Resistor (non-inductive) Current Sense Resistor Power Dissipation Current Sense Resistor Current Sense Resistor Output Fuse Rating

33

V

7.2 IRLZ14PBF

A

0.28

Ω

1.70E-010 0.8 3.50E-009 6.00E-009 1.19E-008 1.902 3.64162787 0.72 1.53E-004 2.16 2.20E-04 33 0.9 680µF/63V 1µF/63V 23.859 0.208

F A C C s W W A H A H V A

2.20E-04 0.084 0.429462 33

F Ω W V

3.55E-008

F

V A

10nF/63V 11.21039121 43 0.429462 0.1 0.1 2.25

Ω Ω W Ω Ω A

100 100 100 100 100 100 100

nF nF nF nF nF nF nF

15

V

Controller Part Values C6 C7 C13 C14 C15 C16 C17

Bypass Capacitors Bypass Capacitors Bypass Capacitors Bypass Capacitors Bypass Capacitors Bypass Capacitors Bypass Capacitors

Auxiliary Power Supply (Voltage Regulator)

46

Gate Drive (Dual Channel Gate Driver from the motor driver circuit) Differential Voltage Sense (optional) Charger Control - IC Setup - Housekeeping and Temperature Sensing U1 C8 fs R8 R7 RP1

TI Timing Capacitor Switching Frequency RSET Oscillator reference resistor - thermistor thermistor emulation Potentiometer

UC3909 1.5 50000 11000 10000 50000

nF Hz Ω Ω Ω

Charger Control - IC Setup - Current Levels R9 R10 R11 R12

OVCTAP set resistor OVCTAP set resistor Trickle Current Limit Set Bulk Current Limit Set

100000 8333.1 1721.74068 5165.22204

Ω Ω Ω Ω

77564.19048 6422.47619 17247.2619 138175.4888

Ω Ω Ω Ω

3078.46354 1.03399E-08 1.03399E-09

Ω F F

1000 3.22E+05 5.73055E-11 4.34048E-08

Hz Ω F F

Charger Control - IC Setup - Voltage Levels R15 R16 R17 R18

Battery Voltage Divider 1% recommended Battery Voltage Divider 1% recommended Battery Voltage Divider 1% recommended Battery Voltage Divider 1% recommended

Charger Control - IC Setup - Current Error Amplifier R14 C11 C12

Current Error Amplifier Compensation Resistor Current Error Amplifier Compensation Capacitor Current Error Amplifier Compensation Capacitor

Charger Control - IC Setup - Voltage Error Amplifier fo R13 C9 C10

Voltage Loop Cross Over Frequency Voltage Error Amplifier Compensation Resistor Voltage Error Amplifier Compensation Capacitor Voltage Error Amplifier Compensation Capacitor

Charge State Controller (N/A - will be used by the micro controller)

47

Appendix C: Software Flowcharts

Software Initialization Start

Setup Stack Pointer

Clear Variables

Setup Timer1 for Fast PWM mode using 16 bit resolution

Setup Timer2 for 1ms interrupt

Set pin PB2 as an output

Clear TCNT1

Clear TCNT2

Load OCR2A with $A5

Call LCD Init

Return from LCD Init

Clear OCR1B

Load OCR1A with $7FE

Setup TCCR2 for Counting up and resetting on match with OCR2A

Setup TCCR1 for Fast PWM mode and 16 bit resolution

Divide internal clock by 32

Start Timer1

Start Timer2

Enable Interrupts

Initialize LCD

Jump to State 01

The SOC calculations are made during an interrupt that occurs every 1ms.

48

Appendix C

Interrupt Service Routine

49

Appendix C

State 1

50

Appendix C

State 2 Jumped to from State 01

No Measure Throttle Potentiometer Voltage

Choose A/D channel AD5

Setup A/D control register to enable A/D, start conversion, and enable interrupt

Disable remaining pins on Port C

Is A/D conversion complete

Yes

Store A/D register as Speed variable

No Measure current

Choose A/D channel AD4

Disable remaining pins on Port C

Stop Timer2

Set PB2 to ‘1’

Setup A/D control register to enable A/D, start conversion, and enable interrupt

Is A/D conversion complete

Yes Store A/D register as Current variable

Set PB2 to ‘0’

Start Timer2

Enable pins on Port C

Jump to State 03

51

Appendix C

State 3

52

Appendix C

State 4 Jumped to from State 01

Jumped to from State 03

Copy Speed Variable to PWM Register

Jump to State 05

State 5

53

Appendix C

State 6 Jumped to from State 05

Update Battery Icon on LCD

Is SOC greater than 80%?

Is SOC between 60% and 79%?

Is SOC between 40% and 59%?

Is SOC between 20% and 39%?

Is SOC between 10% and 19%?

Display Empty Battery Icon

Display 5 Battery Bars

Display 4 Battery Bars

Display 3 Battery Bars

Display 2 Battery Bars

Display 1 Battery Bar

Has 1 second elapsed?

Display “Please Recharge Battery”

Convert Voltage to ASCI value

Display Voltage Value

Convert Current to ASCI value

Display Current Value

Update Speed Icon on LCD

Is Speed Variable Icon =5?

Is Speed Variable Icon =4?

Is Speed Variable Icon =3?

Is Speed Variable Icon =2?

Is Speed Variable Icon =1?

Display 5 Speed Bars

Display 4 Speed Bars

Display 3 Speed Bars

Display 2 Speed Bars

Display 1 Speed Bar

Don’t Display Speed Icon

Jump to State 01

54

Appendix C

LCD Initialization

55