Electric Load Driven Longboard

IN

DEGREE PROJECT TECHNOLOGY, FIRST CYCLE, 15 CREDITS

,

STOCKHOLM SWEDEN 2020

Electric Load Driven

Longboard

JOHAN ANDERSSON

RICKARD HÖGLUND

Electric Load Driven Longboard

JOHAN ANDERSSON RICKARD H ¨OGLUND

Bachelor’s Thesis at ITM Supervisor: Nihad Subasic

Examiner: Nihad Subasic

Abstract

Keywords: mechatronics, longboard, brushless direct cur-rent motor, Hall effect sensor, load cell, Wheatstone bridge, LiPo battery, Arduino, ODrive

This bachelor’s thesis aims to show an extensive overview of all the parts that build up an electric load driven long-board and see if a load controlled longlong-board can be seen as a safe, comfortable and convenient alternative to the more common remote controlled longboard. The thesis will also answer how weight can be measured on a longboard in the most effective way, what the most comfortable riding tech-nique is and what a good motor-battery configuration to be able to travel at 30 km/h and 10 km would be.

The longboard measures the weight distribution with load cells located between the deck and the trucks. An Arduino translates the input from the load cells to a certain speed and then sends it to an ODrive which controls a BLDC mo-tor that is powered by two LiPo batteries.

The results show that a load controlled longboard can very well be seen as a good alternative if right riding technique is used. The best technique is when the longboard acceler-ates when the rider tilts and keeps a constant speed when the rider stands straight. The best way to measure the weight is to fasten the trucks with hinges which lets the load cells register weight without anything interfering. Not all tests could be done because of Covid-19 but a measured top speed of 15 km/h with a high gear ratio is a promising result for the future when more suitable gear ratios will be tested to try to reach the goal of 30 km/h.

Referat

Elektrisk lastdriven longboard

Nyckelord: mekatronik, longboard, borstl¨os likstr¨omsmotor, Hall effekt sensor, lastcell, Wheatstone brygga, LiPo batte-ri, Arduino, ODrive

Det h¨ar kandidatexamensarbetet str¨avar efter att visa en omfattande ¨overblick p˚aalla delar som bygger upp en elekt-risk lastdriven longboard och se om en laststyrd longboard kan ses som ett s¨akert, komfortabelt och beh¨andigt alterna-tiv till den vanligare radiostyrda longboarden via handkon-troll. Det h¨ar arbetet kommer ocks˚a svara p˚a hur vikt kan m¨atas p˚a en longboard p˚a ett s˚a effektivt s¨att som m¨ojligt, vad som ¨ar den mest bekv¨ama ˚akstilen och vad ¨ar en bra motor-batteri konfiguration f¨or att kunna ˚aka i 30 km/h och n˚a 10 km skulle vara.

Longboarden m¨ater viktf¨ordelningen med lastceller som ¨ar placerade mellan br¨adan och truckarna. En Arduino om-vandlar indatan fr˚an lastcellerna till en specifik hastighet som den sedan skickar till en ODrive som kontrollerar en borstl¨os likstr¨omsmotor som i sin tur ¨ar driven av tv˚a LiPo batterier.

Resultaten visar att en laststyrd longboard kan mycket v¨al ses som ett bra alternativ om r¨att ˚akstil anv¨ands. Den b¨asta stilen ¨ar att longboarden accelererar n¨ar ˚akaren lu-tar sig och h˚aller en konstant hastighet n¨ar ˚akaren st˚ar rakt. Det b¨asta s¨attet att m¨ata vikt ¨ar att montera truc-karna p˚a g˚angj¨arn som l˚ater lastcellerna m¨ata vikt utan att n˚agot st¨or. Alla tester kunde inte utf¨oras p˚a grund av Covid-19 men en uppm¨att topphastighet p˚a 15 km/h med en h¨og utv¨axling ¨ar ett lovande resultat f¨or framtiden n¨ar l¨ampligare utv¨axlingar kommer testas f¨or att f¨ors¨oka n˚a m˚alet p˚a 30 km/h.

Acknowledgements

We would like to thank our supervisor Nihad Subasic for his feedback through-out this project and making sure it could all continue despite the situation that was present during this project. We also wish to show gratitude to Seshagopalan Thorapalli Muralidharan for being a huge help when the project became very chal-lenging. He came up with helpful ideas and helped us getting parts manufactured that was needed for the longboard. We would also like to thank Staffan Qvarnstr¨om for ordering parts and teaching us how to solder.

Contents

2.2 Trucks and Wheels . . . 3

2.3 Motor . . . 3

2.4 Hall effect sensor . . . 5

2.5 Battery . . . 6 2.6 Powertrain . . . 7 2.7 Load Cells . . . 7 2.8 Control Unit . . . 8 2.8.1 Microcontroller . . . 8 2.8.2 Motor controller . . . 9 3 Demonstrator 10 3.1 Powertrain . . . 10 3.1.1 Motor . . . 10 3.1.2 Battery . . . 10

3.1.3 Gear Ratio and Wheels . . . 12

3.1.4 ODrive . . . 13

3.1.5 Arduino Uno . . . 13

3.1.6 Programming . . . 13

3.1.7 Gears and Timing Belt . . . 16

3.2 Load Cells . . . 17

3.3 Design . . . 18

3.3.1 Layout . . . 18

3.3.2 Load Cell Mounting . . . 18

4 Results 22

4.1 Load cell placement and weight distribution . . . 22

4.2 Riding Technique . . . 22

4.3 Top Speed . . . 23

4.4 Distance . . . 23

5 Discussion and Conclusion 24 5.1 Discussion . . . 24

5.2 Conclusion . . . 26

6 Recommendations and Future Work 27 Bibliography 28 Appendices 30 A Configuration for the ODrive 32 A.1 Motor configuration . . . 32

A.2 Encoder configuration . . . 33

B Code for the Arduino 34 B.1 Prototype and definitions header file . . . 34

B.2 Code where both HX711 are used . . . 35

B.3 Code where only one HX711 is used . . . 42

C MATLAB code 49

List of Figures

2.1 Parts of the BLDC motor [4]. . . 4

2.2 Motor used to drive the longboard [5]. . . 5

2.3 Hall effect principle with no magnetic field in the left image and with a magnetic field present in the right image [7]. . . 6

2.4 The powertrain setup used on the longboard. Made with Microsoft Pow-erpoint. . . 7

2.5 Load cell amplifier used to connect the load cells with the Arduino [12]. 8 2.6 Load cell used to measure the weight distribution [13]. . . 8

2.7 The ODrive used to control the motor [17]. . . 9

3.1 Battery used to power the longboard [20]. . . 12

3.2 A flowchart of the code. The velocity is measured in Hall effect sensor counts per second. Made with draw.io . . . 15

3.3 Calculated needed distance between the gears. Made with MATLAB. . 16

3.4 300 mm belt [21] used and the ring fitted between the wheel and gear. Made with Solid Edge. . . 16

3.5 Wheatstone bridge configuration with two load cells. Made with Mi-crosoft Powerpoint. . . 17

3.6 The two cases for the batteries, ODrive and Arduino. Made with Solid Edge. . . 18

3.7 Load cell placement of mount type 1, (dimensions not to scale). Made with Microsoft Powerpoint. . . 19

3.8 Deflection of metal plate under the load cells. Made with MATLAB. . . 19

3.9 Mounting type 1 and mounting type 2. Made with Solid Edge. . . 20

3.10 Mounting type 3. The hinges are placed on the near side of the load cell mount. Made with Solid Edge. . . 20

3.11 The motor mount. Made with Solid Edge. . . 21

A.1 The configuration of the motor in ODrive tool. . . 32

A.2 The configuration of the Hall sensor in ODrive tool. . . 33

D.1 The finished longboard . . . 54

D.2 Close up of the powertrain and the load cell mount with its load cells . 55 D.3 The underside of the longboard . . . 56

List of Tables

2.1 Motor specifications . . . 5 2.2 Load cell specifications . . . 8 3.1 Possible battery configurations . . . 11

List of Abbreviations

BLDC - Brushless direct current CPR - Counts per revolution CPU - Central processing unit

GPIO - General-purpose input/output I/O - Input/output

Li-ion - Litium ion LiPo - Lithium polymer MCU - Microcontroller unit mAh - Milliampere hour PLA - Polylactide

RPM - Revolutions per minute

Chapter 1

Introduction

1.1

Background

A longboard is a type of skateboard. Due to its characteristics such as the length and the bigger and softer wheels it rolls a lot better and smoother and is therefore more often used for commuting and downhill riding. A longboard does require a bit of work from the user since the user have to kick to propel it forward. This can get quite exhausting after a while.

Electric longboards already exists on the market. These make it possible to travel without having to kick but the rider has to use a remote controller to pilot it instead. The downsides of this is that the user can lose the controller, it can run out of battery and it can lose connection to the board. To be able to travel on a longboard without having to propel it by kicking or without having to use a remote controller presented itself as an interesting challenge, thus this project was chosen.

1.2

Purpose

The purpose of this thesis is to examine how a load driven electrical longboard should be set up so a person can ride it comfortably and safe. The thesis strive to answer the following questions:

• What is a good motor-battery configuration to have a top speed of 30 km/h and a range of 10 km?

• What is a good way to consistently measure weight on a longboard? • How does a person control the longboard in the most comfortable way? • Can a load controlled longboard be seen as a safe, comfortable and convenient

CHAPTER 1. INTRODUCTION

1.3

Scope

During this project there are time and resource limitations. The course runs throughout the spring term of 2020. Because of this a regular longboard should be purchased second hand and modified to make it electric and load driven. Electrical vehicles such as longboards are defined by The Swedish Transport Agency as bicycles [1]. This means they are not allowed to be driven at speeds exceeding 20 km/h. Since the longboard developed in this project has a goal of cruising at 30 km/h, it will only be used at gated and private areas. Hence this project should also not aim to comply with some of the other laws such as needing headlights, safety reflectors and a bell.

1.4

Method

In order to answer the research questions a regular longboard was purchased second hand. The rest of the components for the project were gathered after research and calculations were done to decide upon what was needed. In the ensuing part of the project designing and constructing took place. The manufactured parts were mostly done by 3D-printing, but also by waterjet cutting.

As the construction was completed and all the functions and features that the board was supposed to have was decided upon, the code was written. It was iterated and updated to include new features several times and was made sure to include safety measures before the board was tested outside with the motor connected to the wheel. When this was completed the different riding techniques, weight measuring components, top speed, and distance could be tested.

Chapter 2

Theory

2.1

Deck

The deck is the part of the longboard where you stand. It is most often made out of wood in several layers. A deck is categorized by its flex which describes how flexible the deck is and how much it bends when you stand on it. The higher the flex the more flexible it is. In this project a lower flex would be preferred to be able to mount the batteries, Arduino Uno and ODrive as good as possible underneath the deck. If the deck flexes too much the cases containing the sensitive components may break due to unexpected loads or due to the fact that they might touch or be dragged along the pavement. The deck can come in many different shapes. This deck will be symmetrical which makes it easy to ride in both directions.

2.2

Trucks and Wheels

To be able to corner and change direction with the longboard special wheel mounts are needed which are called trucks. The trucks translate a leaning motion of the rider to a turning motion of the front and rear axles, thus changing the direction of the board. On this longboard the trucks are 180 mm wide. These also work as mounting point for the motor mount. The wheel type is determined by its diameter and hardness. The wheels for a longboard are generally around 70 mm in diameter, and have a hardness of 80 out of a scale between 75 and 100 where 100 is the hardest.

2.3

Motor

The motor on a longboard needs to be light and strong so a brushless motor is preferred. This is because brushless direct current (BLDC) motors do not use brushes for commutation, instead this is done electronically [2]. Avoiding using brushes means that there is no mechanical contact between the commutator and said brushes, thus the operating life is drastically increased and the BLDC can put

CHAPTER 2. THEORY

out a lot of power and torque relative to its size [3]. The BLDC motor works in such a way that a current is applied to the coils on the stator seen in Fig. 2.1.

Figure 2.1: Parts of the BLDC motor [4].

The coils are energized in pairs to increase efficiency. When energized, they both attract and repel the different poles on the rotor, which makes the rotor spin. To be able to know when and in what sequence to energize these coils, the motor controller will read a sensor [2].

When it comes to deciding which BLDC to use, one will have to check the specifi-cations for the motor. The most important specifispecifi-cations for hobby BLDC motors are described in the equation:

rpm= KV · V, (2.1)

where the KV value describes the linear relation between input voltage and the revolutions per minute (rpm) of the motor. Most motors used to propel longboards generally have a KV rating of slightly under 200 KV. The motor chosen for this project is a Turnigy D5035-125KV Sensored Brushless Motor [5] and can be seen in Fig. 2.2. The specifications are in Table 2.1. The efficiency of this motor is around 70% which means that the output rpm is 70% of what is calculated. It also has a

KV value of 125 KV which generally means that it is more suitable for situations

2.4. HALL EFFECT SENSOR

Figure 2.2: Motor used to drive the longboard [5].

Table 2.1: Motor specifications

Model Turnigy D5035-125KV Sensored Brushless Motor

KV rating 125

Torque constant, KT 0.5 Nm/A

Max power 2000 W

Efficiency ≈70%

Max current 45 A

Shaft 8 mm

Weight 745 g

2.4

Hall effect sensor

As the chosen motor comes with a Hall effect sensor, it makes sense to use this instead of a separate encoder which also can be used. Hall effect sensors work by having a continuous current flow through the Hall element conductor and sensing nearby magnetic fields [6]. When a magnetic field comes close to the sensor a voltage will be generated that is perpendicular to both the current and the field [7]. This is visualised in Fig. 2.3.

CHAPTER 2. THEORY

Figure 2.3: Hall effect principle with no magnetic field in the left image and with

a magnetic field present in the right image [7].

This magnetic field will be created by the energizing of the windings and a voltage will therefore be generated and read by the sensor as the rotor of the BLDC spins [2], [8]. This means that since the position of the rotors are known thanks to the sensors, it is now also known which winding to energize.

2.5

Battery

When deciding upon batteries for a hobby BLDC motor the essential stats comes down to voltage V, battery capacity or electric charge Q, and the continuous dis-charge rate called C rating. Most batteries consists of cells rated to 3.7 V that are connected in series and parallel to get a specific output. Batteries are often named with a S and P value which describes this. When fully charged the cells are 4.2 V and when used they should not surpass 3.5 V to reduce the risk of permanent damage. How many S a battery has describes how many cells are connected in series. The higher the S value, the more volts you can get from the battery. The

P value characterize how many cells are connected in parallel and the higher the P

value, the longer the battery will last before it is empty and the higher the electric charge is. The voltage needed is calculated from equation (2.1) and the required battery capacity in the unit mAh can be calculated by

Q= E

V ·3.6, (2.2)

where the energy E is

E= F · s, (2.3)

where F is the resulting force and s is the wanted distance.

To propel an electric longboard a lot of electric power is needed which means that the batteries required are very powerful and thereby sometimes dangerous to experiment with. For electric longboards most batteries used are either lithium polymer (LiPo) or lithium-ion (Li-ion) [9]. Li-ion are generally safer but more expensive than LiPo batteries. The batteries used in this thesis are LiPo mostly because of the limited budget. Some thought and care need to go into it when handling LiPo batteries.

2.6. POWERTRAIN

One of those things is that a balance charger is needed when charging to prevent overcharging which can cause fires. Charging batteries that are hot to the touch or visibly damaged is ill-advised and exposure to extreme temperatures is discouraged.

2.6

Powertrain

When it comes to propelling an electric longboard there are two established meth-ods. The first is to have a motor inside the wheel, a so called hub motor [10]. Due to availability of a separate BLDC motor, the second option is chosen which is to have the motor mounted to the rear truck and connected to one of the wheels with a belt as seen in Fig. 2.4. A positive thing about this option is that the gear ratio between the motor and the wheel can be chosen to best suit the top speed and acceleration goals. The drivetrain consists of three parts; two gears and a belt. The motor mount sits on one side of the truck and holds the motor. The motor is mounted on the motor mount with four screws in two slots which allows the belt tension to be changed. To prevent slip the belt is put around two pulleys, one on the motor shaft and the other connected to the wheel.

Figure 2.4: The powertrain setup used on the longboard. Made with Microsoft

Powerpoint.

2.7

Load Cells

Compared to the electric longboards on the market today this longboard is con-trolled by the rider’s weight distribution instead of a hand held speed controller. To measure the weight distribution four load cells need to be installed on the longboard between the trucks and the deck. To be able to handle the total weight of the rider two load cells need to be mounted on each truck. A load cell like the one pictured in Fig. 2.6, also known as a single strain gauge, consists of two resistors where one of them may change resistance when the load cell is exposed to external forces. The specifications for the load cell used in this thesis can be found in Table 2.2.

CHAPTER 2. THEORY

Table 2.2: Load cell specifications

maximum measurable weight 50 kg maximum allowed weight 75 kg height 8.7 mm length 33.8 mm width 33.8 mm

With a maximum measurable weight of 50 kg per load cell, each truck can measure 100 kg which is sufficient for the purpose of this thesis. From each load cell there are three wires, a red (E+), a white (O+) and a black (E-) which are connected to the three wires on the other load cell to form a Wheatstone bridge. This ensures that the longboard can measure enough weight without the load cells breaking and it also makes the measuring of the load more accurate [11]. These wires are then connected to the connectors on the load amplifier board, pictured in Fig. 2.5. The load amplifier is then connected to the Arduino Uno via VCC, DAT, CLK and GND.

Figure 2.5: Load cell amplifier used

to connect the load cells with the Ar-duino [12].

Figure 2.6: Load cell used to

mea-sure the weight distribution [13].

2.8

Control Unit

2.8.1 Microcontroller

The microcontroller is used in this project to bridge the gap between the motor controller and the load cells. A microcontroller unit (MCU) is a single-chip com-puter system. There are a lot of different microcontrollers but what they all have in common is that they consist of a central processing unit (CPU), some type of memory and input/output ports (I/O) [14].

2.8. CONTROL UNIT

The Arduino Uno Rev3 is the MCU used in this project to enable the communication between the load cells and the motor controller, in other words the communication between the user and the BLDC motor. The Arduino Uno Rev3 is based on the ATmega328P processor with 16MHz clock speed. It communicates using universal asynchronous receiver-transmitter (UART). With 14 digital and 6 analog inputs and outputs, together with 32 KB of flash memory, it will satisfy the needs for this project [15].

2.8.2 Motor controller

A motor controller is needed to control the BLDC motor. In this case an ODrive is used and can be seen in Fig. 2.7. ODrive is an open source motor controller developed by Oskar Weigl from Sweden [16]. It is specifically designed to make it possible to use inexpensive brushless motors in high performance robotics projects. Normally an encoder is used alongside the ODrive for it to know the current position of the motor, but a Hall effect sensor can be used as a substitute. The ODrive uses the programming language Python and lets the user configure the motor. It also has several interfaces that can be used to communicate with it from other devices, one of which is UART that is conveniently also used by Arduino Uno.

Chapter 3

Demonstrator

There are two major goals set for the performance specifications of the electric longboard:

• The top speed should be 30 km/h.

• The user should be able to travel at least 10 km on one full charge.

3.1

Powertrain

3.1.1 Motor

As the first performance goal with this longboard is to have a top speed of 30 km/h the required voltage needed to get the correct rpm of the motor and the wheel which has a diameter of 74 mm is calculated. The rotational speed of the wheel is

60 2π · v rwheel = 60 2π· 30 3.6 0.074 2 ≈2150.7 rpm (3.1) where v is the speed of the longboard and rwheel is the radius of the wheel. Together

with the KV value and efficiency of the motor the required input voltage can be calculated. With a gear ratio of 1:1 the required voltage is

V = rpm ef f iciency KV = 2150.7 0.7 125 ≈24.6 V. (3.2) 3.1.2 Battery

The other performance goal with the longboard is to be able to go 10 km on one charge. To know how much electric charge is needed the force and energy needs to be calculated. The force is partly the rolling resistance force calculated as

3.1. POWERTRAIN

where Crris the rolling resistance coefficient between the wheel and the surface, m is

the mass of the longboard and rider collectively, and g is the gravitational constant. The mass is set to 90 kg as a maximum value as that is seen as a reasonable number for the user in mind with a good bit of margin to account for the weight of the longboard. The exact value of the rolling resistance coefficient is unknown but 0.04 is a reasonable estimate as the wheels are polyurethane [18]. According to these assumptions the rolling resistance force is

Frr= 0.04 · 90 · 9.81 ≈ 35.3 N. (3.4)

At top speed there will also be some drag that has to be taken into account. The drag force is calculated with the following equation:

FD = 1₂ · ρ · Cd· A · v2 (3.5)

where Cd is the drag coefficient, A is the frontal area, and ρ is the density of the

air. With a frontal area of about 1 m2_{, density of air equal to 1.27 kg/m}2 _{and the}

drag coefficient of 1.2 [19], the drag force is

FD = 1₂ ·1.27 · 1.2 · 1 · 8.3332 ≈52.9 N (3.6)

which inserted in equation (2.3) together with the desired distance s = 10 km results in the energy

E = (Frr+ FD) · s ≈ 882.3 kJ. (3.7)

To get the required capacity equation (2.2) is used:

Q= E V ·3.6 =

882330

24.6 · 3.6 ≈9971.2 mAh. (3.8) These calculations indicates that the battery should have 25 V and about 10 Ah. These calculations are based on the longboard constantly traveling at the top speed of 30 km/h which is not required by the way the goal is set. This means that the required capacity is a bit lower. It was decided that 8 Ah is enough based on the over estimations and battery costs but such battery with the right voltage does not exist. Thus a battery pack has to be created. With the stats needed for the battery decided, the two options presented in Table 3.1 were considered after some research.

Table 3.1: Possible battery configurations

Type In series In parallel Total Total Cells Cost voltage battery capacity

[pieces] [pieces] [V] [mAh] [pieces] [SEK]

LiP o 2 1 29.6 8000 8 1211

CHAPTER 3. DEMONSTRATOR

Since the performance of the configurations are very similar, the big factors consid-ered for the choice was the work that would be required to put everything together and the cost. With this in mind the first configuration with two batteries in series and one in parallel was chosen. This is a Turnigy battery sold by company

Hobbyk-ing [20]. This battery pack has a voltage above the calculated one, which is good

since there will be some energy losses. This battery has a C rating of twelve, which gives a safe continuous discharge rate of

8 Ah · 12C = 96 A (3.9) and with a peak discharge rate of double that. The peak discharge rate is safe to be held for ten seconds [20].

Figure 3.1: Battery used to power the longboard [20].

3.1.3 Gear Ratio and Wheels

With a specific input voltage calculated the desired voltage and top speed can still be changed by changing the wheel diameter and gear ratio between the motor and wheel. The most common wheel size is around 70 mm in diameter but they can be much larger. The gear ratio may also differentiate. By increasing the gear ratio the acceleration will increase and the top speed will decrease. The acceleration is given by: m · a= Mv rwheel = Kt· I rwheel . (3.10)

With a gear ration of 1:1 and maximum current of 45 A the theoretical acceleration will be:

a= 0.5 · 45

90 · 0.037 ≈6.8 m/s2. (3.11) Such an acceleration is far too much for a longboard, so the acceleration is set to a more reasonable value in the code. In the experiment phase of the project three different sizes of the gear on the motor will be tested to see the difference.

3.1. POWERTRAIN

3.1.4 ODrive

The main purpose of the ODrive is to control the electric motor. As it can work together with an Arduino, and exists in 56 V version with a peak current of 120 A per motor, it is well suited for this project. The batteries are directly connected to the ODrive via XT90 connectors and the ODrive is connected to the motor via three phase wires. The ODrive is also connected to the Arduino via three wires, two from the digital pins on the Arduino to GPIO on the ODrive, and one between GND on the Arduino and GND on the ODrive. To protect the ODrive the motor controller is encased in a 3D printed case that has three fans attached to it to prevent the ODrive from overheating. The ODrive may get very hot because of the high current running through it. When the ODrive is connected to the motor and batteries, the configurations could take place. This is done in two parts, one for the motor and one for the Hall sensor. For the motor it was mostly setting parameters that define the way it behaves. This includes, amongst other things, velocity limit, current limit and tolerances. The full configuration can be seen in Appendix A.1. For the Hall sensor the calibration included finding out the amount of pole pairs the BLDC motor has. By applying a low current one could spin the motor and feel the resistance when passing the poles. By this method the pairs were found out to be seven. With this information the counts per revolution (CPR) which is the encoder resolution could be set. The full encoder configuration can be seen in Appendix A.2.

3.1.5 Arduino Uno

The main purpose of the Arduino is to translate inputs from both load cell amplifiers to a signal fed to the ODrive that describes what speed the motor should be rotating at. To be able to take inputs from the load cell amplifiers and give inputs to the ODrive two special libraries need to be downloaded to the Arduino. In that way new specific commands can be used to both take information from the amplifiers and the ODrive but also send information to the ODrive. In this case the Arduino tells the ODrive what speed the motor should be running in Hall effect sensor counts per second. The Arduino is powered by a separate battery pack with six AA batteries connected in series. The nine-volt battery pack is connected to the power port via an aircraft inspired switch.

3.1.6 Programming

Two different types of riding techniques were tested. Fig. 3.2 shows a compre-hensive overview of the first riding technique in a flowchart, which is the one that seemed as most intuitive before the test rides. The basic idea is that the longboard is controlled by tilting. The longboard accelerates in the direction the rider tilts or keeps a constant speed if the rider stands straight except for when the speed is close to zero, then it will accelerate either forwards or backwards to keep the velocity at zero. This is done in order to make sure the longboard can stay still without having

CHAPTER 3. DEMONSTRATOR

the user needing to continuously shift their weight, and therefore enables the user to relax. This is a valuable feature when the user is at traffic lights where unwanted movement might be devastating and when trying to stand still on a surface that is not level. The code will constantly check if the load cells register enough weight and if they do not the longboard will come to a full stop quickly without locking the wheels to ensure the safety of the board, other pedestrians and property. The voltage of the LiPo batteries is also constantly monitored. If the voltage gets close to what is set as drained, the top speed will be reduced to notify the rider. When it actually is empty the longboard will come to a stop and the Arduino is put to sleep so no more riding can be done. The entire code can be found in Appendix B.2 together with a header file in Appendix B.1 used to clean up the main program. The flowchart of the second riding technique is almost identical to the one cor-responding to the first riding technique. The difference is that instead of a tilt resulting in a constant acceleration, a tilt results in a specific set speed. Depending on how much the rider tilts the speed changes to the speed equivalent of a specific weight distribution. The more the rider tilts in one direction the faster the long-board goes in that direction and if the rider is standing straight the velocity is set to zero.

3.1. POWERTRAIN

Figure 3.2: A flowchart of the code. The velocity is measured in Hall effect sensor

CHAPTER 3. DEMONSTRATOR

3.1.7 Gears and Timing Belt

The gear mounted on the wheel is a HTD 5M gear with 36 teeth and has a diameter of 56 mm. The 5M means that the pitch, or the distance between each tooth, is 5 mm. The side of the gear is cone shaped but to be able to fit it to the wheel that has a flat side, a 3D printed ring, as seen in Fig. 3.4 is placed between them. The HTD 5M gears mounted on the motor have three different sizes. The smallest has 15 teeth and is 20 mm in diameter, the largest has the same size as the one mounted on the wheel and the third gear has teeth and a diameter somewhere between the smallest and largest. The motor mount is designed to have a distance between the gears equal to 75-100 mm which is enough with three different sized HTD 5M belts with lengths 300 mm, 330 mm and 350 mm. This is shown below in Fig. 3.3. The calculations where made with MATLAB and the code can be found in Appendix C.

Figure 3.3: Calculated needed distance between the gears. Made with MATLAB.

Figure 3.4: 300 mm belt [21] used and the ring fitted between the wheel and gear.

3.2. LOAD CELLS

3.2

Load Cells

The pair of load cells on each truck is connected as a Wheatstone bridge which is shown below in Fig. 3.5.

Figure 3.5: Wheatstone bridge configuration with two load cells. Made with

Microsoft Powerpoint.

The Wheatestone Bridge is then connected to the load cell amplifier which is con-nected to the Arduino via GND, 5V, AREF, and two digital pins. The Arduino is in other words connected to two amplifiers each giving its own weight input.

CHAPTER 3. DEMONSTRATOR

3.3

Design

The longboard weighs 7.6 kg, is 99 cm long and 27.5 cm wide. Pictures of the completed longboard can be seen in Appendix D.

3.3.1 Layout

The batteries and the control unit are placed as close as possible to the trucks where the deck is least bent. Everything except the load cells will be encased in two separate cases, as seen in Fig. 3.6, to protect them from gravel, stone chips, water or other materials or liquids that may cause damage. The batteries are placed in the case shown to the left which sits behind the truck in the front. The other case encloses the ODrive, the Arduino, the battery pack for the Ardunio and the load cell amplifiers. That case is placed in the rear as close to the motor as possible where the airflow is cleaner [22]. The case is placed there because the ODrive runs the largest risk of overheating and therefore needs as much airflow through the case as possible. As seen in Fig. 3.6, the case enclosing the ODrive has an air intake that is placed on the side the ODrive lies, supplying fresh air that runs directly over the fans that cools the ODrive. Each case is 3D printed with polylactide (PLA) plastic as two separate pieces and are then glued together with superglue. The cases are mounted to the deck with M4 screws.

Figure 3.6: The two cases for the batteries, ODrive and Arduino. Made with Solid

Edge.

3.3.2 Load Cell Mounting

As it was difficult to be sure how the designs of the load cell mounting would work in actuality, three different versions were created that was thought to work, although some were believed to work better than the others. What can be called load cell mounting type 1 were designed so that the load cells are placed on a metal plate that is wider than the truck as the load cells would not fit otherwise. To measure the weight distribution the load cells are placed between the trucks and the deck as seen in Fig. 3.7.

3.3. DESIGN

Figure 3.7: Load cell placement of mount type 1, (dimensions not to scale). Made

with Microsoft Powerpoint.

To make sure the longboard remain stable the load cells are mounted next to an equally thick but bigger foam plate. To make the load cells take up most of the load they are placed on the inside of the trucks. The load cells sit on a thin metal plate that needs to be just thick enough to not bend. If it bends the load cells will not work properly. To calculate the bending of the metal plate a formula from solid mechanics from KTH is used [23]. Elementary case 3 gives the deflection:

δ(ξ) = Ql 3

24EI(ξ4−4ξ − 3) (3.12) Where Q [N] is the force, E [N/m2_{] the elastic modulus, I [m}4_{] the second moment}

of area, l [m] the length of the plate and ξ the position where zero is at the end. A metal plate with a thickness of 4 mm gives a deflection of just over one tenth of a millimeter which is acceptable, see Fig 3.8.

CHAPTER 3. DEMONSTRATOR

Another way to mount the load cells is to place them directly above the trucks between them and the deck. This one is referred to as type 2. A CAD-drawing of both mounting types can be seen in Fig. 3.9.

Figure 3.9: Mounting type 1 and mounting type 2. Made with Solid Edge.

A third way to mount the load cells is to mount the trucks to the deck with hinges. In that way the trucks are securely fastened in most directions on the outer edge of the truck but allows rotation. By placing the load cells on the other side of the truck towards the middle between the truck and the deck it is made sure that the load cells will take up all the weight placed upon the board. This makes sure that no weight is taken up by something else than the load cells which could otherwise happen when the mount and deck is screwed together like on type 1 and type 2. Load cell mounting type 3 can be seen in Fig. 3.10. All three mounting types will be tested and the one that yields the best results, in other words, the one that makes the load cells measure the weight distribution most accurately, is used in the final product.

Figure 3.10: Mounting type 3. The hinges are placed on the near side of the load

3.3. DESIGN

3.3.3 Motor Mount

The motor mount is manufactured by waterjet cutting. The idea was to use stainless steel, but as the two possible options were regular steel and aluminium the decision fell upon aluminium. This was due to the fact that aluminium is more resistant to rust than regular steel which is an important aspect to consider since the longboard might run through water if it starts raining while out. Aluminium is also a lot lighter which is preferred because a heavy longboard is difficult to carry, and takes more power to propel. The downside of choosing aluminium is that it gets scratched and damaged aesthetically more easily, but as aesthetics is a lower priority than function for this project this did not matter. The motor mount part can be seen in Fig. 3.11. The mount is clamped to the rear truck with an M5 screw. The hole opposite to the screw head is threaded, so there’s no need to use a nut to get the clamping effect desired. The BLDC motor is attached to the motor mount on the right side in the elongated holes which makes it possible to vary the distance between the rear axle and the motor shaft. The shortest distance is 7 cm and the longest is 11 cm which is a larger span than calculated but is needed so the belt may be tightened properly.

Chapter 4

Results

4.1

Load cell placement and weight distribution

All load cell mounting types were tested with 3D printed mounts. Load cell mount

type 1 did not work at all. The load cells registered 0 kg. Load cell mount type 2 worked fairly well. They could to some extent measure the weight distribution

but the trucks had to be mounted very loose for the load cells to work. Load cell mount type 3 worked satisfactorily. They could measure the weight distribution far better than the other two types. However when the rider picked up the longboard the trucks rotated freely and partially damaged the thin wires connected to the load cells in the front. A piece of duct tape was used on the the rear truck to prevent this from happening on the rear load cells in the following test runs. To perform the other tests the code was slightly modified to only have inputs from the fully working load cells in the rear. The second code can be found in Appendix B.3. It was difficult for the load cell to continuously measure the weight distribution accurately when going at higher speeds which made it difficult to follow the second riding technique. Furthermore, a problem that appeared for all mounting types were that the load cells made a small dent in the deck when under load which made it more difficult for them to measure accurately.

4.2

Riding Technique

The most comfortable riding technique was the first one where the longboard ac-celerates when the rider tilts and keeps a constant speed when the rider stands straight. It was reasonably easy to accelerate and keep a constant speed for both a new and an experienced rider. The other riding technique was fairly difficult to follow as previously said. It was uncomfortable to keep tilting when riding and because the load cells were not that accurate it was hard to control the longboard.

4.3. TOP SPEED

4.3

Top Speed

The only available gear ratio was 2.4 with a 36 tooth gear on the wheel and a smaller 15 tooth gear on the motor shaft. With that configuration and almost fully charged batteries the top speed was around 15 km/h.

4.4

Distance

When calculating the electric charge needed to be able to travel 10 km the top speed was used. It was not possible to ride at full speed during the entire test run because there were no roads empty enough to ride at full speed continuously. According to the measurements the distance traveled was roughly 12 km. The batteries were not fully discharged.

Chapter 5

Discussion and Conclusion

5.1

Discussion

Load cell mount type 1 did not work at all. The load cells were placed too close to the truck for them to take up any weight. Instead all the weight was taken up by the trucks and the load cells registered 0 kg. When weight is applied on the longboard the deck flex and the trucks tilt slightly inwards making the wheelbase a slight bit longer. This results in that the load cell mounts together with the load cells also tilt and the load cells will not be able to register any difference in applied weight.

Load cell mount type 2 worked fairly well but for them to work properly and take up all the weight the load cells need to stick up above the edge of the load cell mount which results in an unstable mounting of the entire truck.

The best load cell mount was type 3. They worked satisfactorily as well but this time the trucks could be screwed tightly to the deck which results in a much more stable ride. They were the best because the load cells were completely free from screws or any other objects that could have interfered with the measuring of the weight. One drawback with this mount is that there needs to be some kind of stop for them so when the rider picks up the longboard the trucks will not rotate more than a few degrees. The thin cables connected to the load cells will otherwise get damaged. This mount is used in the final product.

As previously mentioned the load cells made a small dent in the deck which may have made it more difficult for them to measure weight accurately. To prevent this from happening small thin metal plates were placed where the load cells hit the deck making sure the deck is not damaged and the reading keeps being accurate.

5.1. DISCUSSION

The first riding technique was deemed superior where the longboard accelerates when the rider tilts and keeps a constant speed when the rider stands straight. This riding technique was fairly easy to get used to and felt intuitive as previously thought. The load cells are good enough to note if the weight distribution is more than 55/45 in any direction. The acceleration and retardation were good but some-times there were some delay between the rider tilting and the motor changing speed. This is assumed to be because of a communication issue between the Arduino and ODrive. The speed of communication that is required by the ODrive is possible with an Arduino Uno but it is not reliable, thus some data may be lost sometimes. This could be avoided by using hardware communication instead of the software communication used in this project.

When it comes to safety the longboard is fairly good. As the top speed is not as high as previously expected, it is quite easy to jump off the longboard at speed if something goes wrong while riding. One thing that was noticed during the top speed test was that it is difficult to lean back to slow down the longboard when riding at a high speed and the ride gets moderately unstable. This is believed to partly be because of the sometimes recurring communication issue between the ODrive and Arduino.

When jumping off the longboard at speed there seems to be some generative braking which results in a lower measured voltage and maybe a charge going back into the batteries. The ODrive register an unacceptable low voltage and the Arduino goes into sleep mode as was coded. When that happens the batteries and Arduino need to be unplugged and then plugged in again to restart the ODrive and Arduino. This was to a large extent solved by having the Arduino going into sleep mode when the voltage was measured to just slightly under 28 V, and not react at all when mea-suring excessively low voltages.

One thing to note is that the measured top speed was 15 km/h when a gear ratio of 2.4 was used. This speed is slightly higher than expected as a top speed of 30 km/h when using a gear ratio of 1:1 should result in a speed of just under 13 km/h with the gear ratio 2.4. But as drag increases by v2 _{as seen in equation (3.5) it makes}

sense that it’s not linear.

As previously said the distance covered by one charge was about 12 km which is quite difficult to interpret. The top speed was lower than what was used in the calculations, the batteries were only 8000 mAh instead of the calculated 10000 mAh and the speed was not 15 km/h the whole test run. With some backwards calculations with these parameters the distance a rider should be able to travel on one full charge is about 17.5 km. This distance is a lot larger than actually covered, but 12 km may be reasonable because the roads were neither smooth, which impacts the friction, nor flat and the batteries were not fully discharged.

CHAPTER 5. DISCUSSION AND CONCLUSION

5.2

Conclusion

The research questions this thesis strives to answer are as previously mentioned: • What is a good motor-battery configuration to have a top speed of 30 km/h

and a range of 10 km?

• What is a good way to consistently measure weight on a longboard? • How does a person control the longboard in the most comfortable way? • Can a load controlled longboard be seen as a safe, comfortable and convenient

alternative to a remote controlled longboard?

For a rider to be able to travel at 30 km/h and have a range of 10 km the longboard needs two 4S LiPo batteries connected in series powering a 125 KV BLDC electric motor with 29.6 V and 8000 mAh. The top speed was 15 km/h but with a higher gear ratio than what was used in the calculations. The distance covered on almost one full charge was roughly 12 km.

The best way to measure weight consistently on a longboard is to fasten the truck with hinges on the outer side and place the load cells on the inner side of the truck. This makes sure that nothing will interfere with the load cells.

The most comfortable way for a person to control the longboard is to control the acceleration and braking by tilting and to keep a constant speed when standing straight. Even though the load cells were not perfectly accurate this riding tech-nique worked fine.

The load controlled longboard can be seen as a good alternative to the more common remote controlled longboard because it was comfortable and convenient to pilot, and all the safety features worked well. One does not miss the remote control.

Chapter 6

Recommendations and Future Work

Due to the current situation with the Covid-19 pandemic not all parts needed were manufactured. Because of that the different gear ratios could not be tested. This is something that can be done in the future to reach the desired top speed by get-ting a 20-24 tooth gear and a 36 tooth with matching belts and also manufacturing keyways in the gears so they can be mounted on the motor shaft.

A big recommendation would be to edit the Arduino code so that it uses hardware communication instead of software. Since the Arduino Uno only has one hardware serial port, problems can appear when writing the code as both the ODrive and computer would use the same port. But having a finished code and then rework it is expected to make the longboard more responsive. This was realized too late in the project and could not be implemented due to the time constraint.

An useful feature would be to implement a display that shows the battery percent-age on the longboard so the user can see when the batteries should be charged. The display should be connected to the Arduino.

The current cover for the batteries is acceptable and works, but a solution that is more simple to open in order to get the batteries out to charge them could make it more convenient. Using hinges on one side and something to hold it closed that can be opened easily on the other would be an improvement.

Bibliography

[1] Cykel - Transportstyrelsen, Transportstyrelsen.se, 2013. [Online]. Available: https : / / www . transportstyrelsen . se / sv / vagtrafik / Fordon / Fordonsregler/Cykel/(visited on 01/27/2020).

[2] P. Yedamale, “Brushless DC (BLDC) Motor Fundamentals”, Microchip

Tech-nology Inc, vol. 20, 2003.

[3] S. Rambabu, “Modeling and control of a brushless dc motor”, M.S. thesis, National Institute of Technology Rourkela, Rourkela, Odisha, India, 2007. [Online]. Available: http://ethesis.nitrkl.ac.in/4304/ (visited on 03/12/2020).

[4] How Brushless Motor and ESC Work - HowToMechatronics, HowToMecha-tronics.com. [Online]. Available: https : / / howtomechatronics . com / how- it- works/how- brushless- motor- and- esc- work/ (visited on 02/16/2020).

[5] Turnigy D5035-125KV Sensored Brushless Motor, Hobbyking.com. [Online]. Available: https://hobbyking.com/en_us/dt6376-14p-sensored-motor-125kv.html?___store=en_us(visited on 02/13/2020).

[6] Z. Zafirakis, “Development of a testing assembly for powertrain speed sensors”, Master’s thesis, KTH, School of Electrical Engineering (EES), Stockholm, Sweden, 2016. [Online]. Available: http://urn.kb.se/resolve?urn= urn:nbn:se:kth:diva-194222.

[7] Hall Effect Sensing and Application, English, Honeywell Inc., Freeport IL, USA. [Online]. Available: https://sensing.honeywell.com/honeywell- sensing-sensors-magnetoresistive-hall-effect-applications-005715-2-en3.pdf(visited on 03/29/2020).

[8] M. Edling Huv´en, “Electric Motor Control”, Master’s thesis, KTH, Machine Design (Dept.), Stockholm, Sweden, 2010. [Online]. Available: http://urn. kb.se/resolve?urn=urn:nbn:se:kth:diva-49587.

[9] S. Lachlan, “The design, manufacturing and testing of a tracked electric mountain board”, Honours thesis, The University of Queensland, 2016. doi: https://doi.org/10.14264/uql.2017.224.

BIBLIOGRAPHY

[10] L. S¨odergren, “Electric longboard : A dual-purpose personal vehicle”, B.S. thesis, KTH, Machine Design (Dept.), 2018. [Online]. Available: http:// urn.kb.se/resolve?urn=urn:nbn:se:kth:diva- 230590 (visited on 03/12/2020).

[11] Wheatstone Bridge Circuit and Theory of Operation, Electronics-Tutorials.ws. [Online]. Available: https://www.electronics-tutorials.ws/blog/ wheatstone-bridge.html(visited on 02/14/2020).

[12] K¨op F¨orst¨arkare f¨or lastceller - HX711 till r¨att pris @ Electrokit, Electrokit.com. [Online]. Available: https://www.electrokit.com/produkt/forstarkare-for-lastceller-hx711/(visited on 02/16/2020).

[13] K¨op Lastcell 50Kg till r¨att pris @ Electrokit, Electrokit.com. [Online]. Avail-able: https://www.electrokit.com/produkt/lastcell- 50kg/ (visited on 02/16/2020).

[14] D. E. Bolanakis, Microcontroller Education: Do it Yourself, Reinvent the

Wheel, Code to Learn. Morgan & Claypool, 2017, p. 11. doi: 10 . 2200 /

S00802ED1V01Y201709MEC009. [Online]. Available: https://ieeexplore. ieee.org/document/8106902.

[15] Arduino Uno Rev3, Arduino. [Online]. Available: https://store.arduino. cc/arduino-uno-rev3(visited on 05/06/2020).

[16] ODrive, ODrive. [Online]. Available: https://odriverobotics.com/ (visited on 05/06/2020).

[17] ODrive - Highperformance motor control, Hackaday. [Online]. Available: https: / / hackaday . io / project / 11583 odrive high performance -motor-control(visited on 05/09/2020).

[18] D. Lippert and J. Spektor, Rolling Resistance and Industrial Wheels, Hamil-ton Caster & Mfg. Co., HamilHamil-ton OH, USA. [Online]. Available: http:// www . hamiltoncaster . com / Portals / 0 / blog / White % 20Paper % 20Rolling%20Resistance.pdf(visited on 02/20/2020).

[19] A. S. Al-Obaidi and M. Koo, “Calculation of Aerodynamic Drag of Hu-man Being in Various Positions”, Presented at The 1st Eng.

Undergradu-ate Res. Catalyst Conf., Jul. 2013. [Online]. Available: https : / / www .

researchgate . net / publication / 321866543 _ Calculation _ of _ Aerodynamic_Drag_of_Human_Being_in_Various_Positions (vis-ited on 02/16/2020).

[20] Turnigy High Capacity 8000mAh 4S 12C Lipo Pack w/XT90, Hobbyking.com. [Online]. Available: https://hobbyking.com/en_us/turnigy-high-capacity - battery - 8000mah - 4s - 12c - drone - lipo - pack - xt90 . html ? queryID = c346abf36360a2013613a1ad842255c0 & objectID = 78430&indexName=hbk_live_magento_en_us_products&___store= en_us(visited on 02/11/2020).

BIBLIOGRAPHY

[21] RS PRO, Timing Belt, 60 Teeth, 300mm Length X 15 mm Width, RS Com-ponents. [Online]. Available: https: / / se . rs- online . com / web / p/ products/1755198/(visited on 03/29/2020).

[22] N. Sriram, D. P. Chandar, V. A. Subramanian, and M. Ganesh, “Design op-portunities in electric skateboard”, in 2017 International Conference on

In-telligent Computing, Instrumentation and Control Technologies (ICICICT),

Jul. 2017, pp. 1075–1078. doi: 10.1109/ICICICT1.2017.8342718.

[23] B. Broberg et al., “Elementarfall”, in Handbok och formelsamling i H˚allfasthetsl¨ara,

B. Alfredsson, Ed. Stockholm, Sweden: Department of Solid Mechanics, KTH, 2016, p. 356.

APPENDIX A. CONFIGURATION FOR THE ODRIVE

Appendix A

Configuration for the ODrive

A.1

Motor configuration

A.2. ENCODER CONFIGURATION

A.2

Encoder configuration

Appendix B

Code for the Arduino

B.1

Prototype and definitions header file

2 * Authors: Johan Andersson & Rickard H¨oglund 3 * Date: 2020-05

4 * File for #define's and function prototypes 5 * to ELCBoard.ino

6 */

7 //===Defining pins and constants===

8 #define CLK A0 // clock pin to the

load cell amp

9 #define DOUT1 A1 // Data pin to the

first load cell amp

10 #define DOUT2 A2 // Data pin to the

second load cell amp

11 #define calibration_factor 5/100000 // Calibration factor 12 #define TARE_TIMEOUT_SECONDS 4

13 #define CHANNEL_COUNT sizeof(DOUTS)/sizeof(byte)

14 #define sei() //Enables interrupts

15 16 //===Function prototypes=== 17 void goingToSleep(); 18 void tare(); 19 void readLoadCells(); 20 void leanForward(); 21 void leanBackward(); 22 void standingUpright(); 23 void enoughWeight(); 24 void checkVoltage(); 25 void batteryLow();

B.2. CODE WHERE BOTH HX711 ARE USED 26 void batteryEmpty();

B.2

Code where both HX711 are used

1 /* File name: ELCBoard.ino (ELCBoard - Electric Load Controlled Longboard)

2 * Project name: Electric Load Driven Longboard 3 * Authors: Johan Andersson & Rickard H¨oglund 4 * Date: 2020-05

5 * School: KTH Royal Institute of Technology 6 * Course: MF133x

7 * Description: This code controls a longboard by load distribution

8 * on the deck, thus not needing a remote control 9 * Hardware:

10 * MCU: ATmega328

11 * Board: Arduino Uno Rev3 12 * External hardware:

13 * Motor controller: ODrive v3.6 56V 14 * Load cell amplifiers: SparkFun HX711 15 */ 16 17 //=======Include libraries======= 18 #include "HX711-multi.h" 19 #include "HX711.h" 20 #include <ODriveArduino.h> 21 #include <SoftwareSerial.h> 22 #include <avr/sleep.h> 23 #include <avr/interrupt.h>

24 #include "prototypesAndDefs.h" //.h-file with prototypes and #defines

25

26 //===Connects serial to ODrive and make ODrive object=== 27 SoftwareSerial odrive_serial(8, 9);

28 ODriveArduino odrive(odrive_serial);

29

30 //=======Declaring constants======= 31 byte DOUTS[2] = {DOUT1, DOUT2};

32 long int results[CHANNEL_COUNT];

33 HX711MULTI scales(CHANNEL_COUNT, DOUTS, CLK);

34 int gain = 20;

APPENDIX B. CODE FOR THE ARDUINO 36 int max_velocity = 2000; 37 int min_weight = 20; 38 float weight; 39 float weightFront; 40 float weightRear; 41 float volt;

42 bool start = false;

43

44 //=======Setup is run once======= 45 void setup() {

46 odrive_serial.begin(115200); //Establish communication between ODrive and Arduino

47 Serial.begin(115200); //Comment out when not

connected to PC

48 Serial.flush(); //Make sure all data in

buffer is sent

49 tare(); //Tare load cells

50

51 odrive.SetVelocity(0, velocity); //Make sure velocity is 0

52

53 pinMode(LED_BUILTIN, OUTPUT); //Use led on pin 13 to see when Arduino is asleep

54 digitalWrite(LED_BUILTIN, HIGH); //Turn LED on to indicate Arduino is awake

55 }

56

57 //=======Loop that will run continuously======= 58 void loop() {

59

60 //Check if there is enough power in the LiPo batteries supplying the ODrive

61 checkVoltage(); //Call function checkVoltage()

62 if (volt < 28.5 && volt >= 28) { //When battery is getting close to empty:

63 batteryLow(); //Call function to

make board go slower 64 }

65 else if (volt < 28 && volt > 26) { //If battery is empty: 66 batteryEmpty(); //Come to a full stop

and put Arduino to sleep (safety having ">26" as a

67 } //single faulty

B.2. CODE WHERE BOTH HX711 ARE USED 68

69 //Read the load cells

70 readLoadCells(); //Returns array, results[0] is front and results[1] is back of board 71

72 if (start == false) { //Replays until minimum weight is reached

73 if (weight > min_weight) { //Check if weight requirement is met

74 delay(3000); //If it is, wait 3 seconds

so the person can get up fully on the longboard 75 readLoadCells(); //Read load cells again 76 start = true; //Set start to true so rest

of code can run

77 }

78 }

79

80 //In this loop the commands that runs the motor are called depending on weight distribution

81 if (start == true) { //If minimum weight is reached, this block of code runs

82 if (weightFront > 0.55 * weight && weightRear < 0.45* weight) { //If leaning forward

83 leanForward(); //Call function

leanForward to accelerate

84 }

85 else if (weightFront < 0.45 * weight && weightRear > 0.55*weight) { //If leaning backward

86 leanBackward(); //Call function

leanBackward to brake or reverse

87 }

88 else { //If standing

straight

89 standingUpright(); //Call function standingUpright

90 }

91 enoughWeight(); //Check if weight

still is enough, if not: stop 92 }

93 }

94

95 //=======Declaring various functions used in the code=======

96 void tare() { //Function to tare

APPENDIX B. CODE FOR THE ARDUINO 97 bool tareSuccessful = false;

98 unsigned long tareStartTime = millis();

99 while (!tareSuccessful && millis() < (tareStartTime + TARE_TIMEOUT_SECONDS * 1000)) {

100 tareSuccessful = scales.tare(20, 10000); //reject 'tare' if still ringing

101 }

102 }

103

104 void readLoadCells() { //Function

reading load cells

105 scales.read(results); //Gets a

reading

106 weightFront = results[0] * calibration_factor; //Set value of front load cells

107 weightRear = results[1] * calibration_factor; //Set value of rear load cells

108 weight = weightFront + weightRear; //Calculate total weight

109 }

110

111 void leanForward() { //Function when leaning forward

112 if (velocity < max_velocity) { //Only increase velocity if it's not already max

113 velocity = velocity + gain; //Increase velocity 114 odrive.SetVelocity(0, velocity); //Send command to

ODrive with the new velocity 115 }

116 }

117

118 void leanBackward() { //Function when leaning backward

119 if (velocity > -max_velocity) { //Only reduce velocity if it's not already max

120 velocity = velocity - gain; //Reduce velocity 121 odrive.SetVelocity(0, velocity); //Send command to

ODrive with the new velocity 122 }

123 }

124

125 //Function for when standing upright, if slow keep velocity 0 so the board doesn't roll. If at higher speed, do nothing

B.2. CODE WHERE BOTH HX711 ARE USED 126 void standingUpright() {

127 if (velocity > 0 && velocity < 150) { //If going forward slowly

128 leanBackward(); //Call function

leanBackward to stop

129 delay(100); //100ms delay so

as to not get too jerky movements 130 }

131 if (velocity < 0 && velocity > -150) { //If reversing slowly

132 leanForward(); //Call function

leanForward to stop

133 delay(100); //100ms delay so

as to not get too jerky movements 134 }

135 }

136

137 void enoughWeight() { //Function to check if weight is enough

138 if (weight < min_weight) { //If weight on board now is less than minimum weight

139 if (velocity > 0) { //If going forward, slow down

140 while (velocity != 0) { //All the way to zero 141 leanBackward(); //By calling function

leanBackward

142 delay(10); //Delay of 10ms so as to

not lock and tear wheels

143 }

144 }

145 if (velocity < 0) { //If reversing, slow down 146 while (velocity != 0) { //All the way to zero 147 leanForward(); //By calling function

leanForward

148 delay(10); //Delay of 10ms so as to

not lock and tear wheels

149 }

150 }

151

152 delay(2000); //Wait for two seconds

153 Serial.flush(); //Waits for outgoing

serial buffer to empty

154 while (Serial.available()) { //This while loop waits for the incoming buffer to empty

APPENDIX B. CODE FOR THE ARDUINO

155 Serial.read();

156 }

157 tare(); //Tare load cells

158 start = false; //Set start to false 159 }

160 }

161

162 void checkVoltage() { //Function to check voltage supplied to ODrive

163 Serial.flush(); //Waits for

outgoing serial buffer to empty

164 while (Serial.available()) { //This while loop waits for the incoming buffer to empty,

165 Serial.read(); // otherwise

problems reading voltage comes due to Arduino Uno's

166 } // instability

of using SoftwareSerial at this high baud-rate

167 odrive_serial.write("r vbus_voltage\n"); //Read voltage supplied to ODrive and write to serial

168 volt = odrive.readFloat(); //Assign to float

169 }

170

171 void batteryLow() { //Function to slow down when battery is low

172 while (velocity > 1000) { //If going forwards quickly, slow down to 1000counts/s

173 leanBackward(); //Slowing down

174 delay(200); //Delay 200ms each iteration,

ramp down slowly 175 }

176 while (velocity < -1000) { //If reversing quickly, slow down to 1000counts/s

177 leanForward(); //Slowing down

178 delay(200); //Delay 200ms each iteration,

ramp down slowly 179 }

180 max_velocity = 1000; //Set new top speed to 1000 counts/s

181 }

182

183 void batteryEmpty() { //Function to stop when battery is empty

B.2. CODE WHERE BOTH HX711 ARE USED

185 while (velocity != 0) { //slow down all the way to zero 186 leanBackward();

187 delay(100); //Delay 100ms, ramp down slowly

188 }

189 }

190 if (velocity < 0) { //If reversing, 191 while (velocity != 0) { //slow down to zero 192 leanForward();

193 delay(100); //Delay 100ms, ramp down slowly

194 }

195 }

196 delay(5000); //After zero velocity is reached, wait 5 seconds

197 goingToSleep(); //Put Arduino to sleep 198 }

199

200 void goingToSleep() { //Function putting Arduino to sleep

201 set_sleep_mode(SLEEP_MODE_PWR_DOWN); //Setting the sleep mode, in our case full sleep

202 sleep_enable(); //Enable sleep mode

203 sei(); //Setting interrupts

204 digitalWrite(LED_BUILTIN, LOW); //turning LED off

205 delay(1000); //Wait to allow the

led to be turned off before going to sleep

206 sleep_cpu(); //Put Arduino to

sleep 207 }

APPENDIX B. CODE FOR THE ARDUINO

B.3

Code where only one HX711 is used

1 /* File name: ELCBoardRearCells.ino (ELCBoard - Electric Load Controlled Longboard)

2 * Project name: Electric Load Driven Longboard 3 * Authors: Johan Andersson & Rickard H¨oglund 4 * Date: 2020-05

5 * School: KTH Royal Institute of Technology 6 * Course: MF133x

7 * Description: This code controls a longboard by load distribution

8 * on the deck, thus not needing a remote control 9 * Hardware:

10 * MCU: ATmega328

11 * Board: Arduino Uno Rev3 12 * External hardware:

13 * Motor controller: ODrive v3.6 56V 14 * Load cell amplifiers: SparkFun HX711 15 */ 16 17 //=======Include libraries======= 18 #include "HX711-multi.h" 19 #include "HX711.h" 20 #include <ODriveArduino.h> 21 #include <SoftwareSerial.h> 22 #include <avr/sleep.h> 23 #include <avr/interrupt.h>

24 #include "prototypesAndDefs.h" //.h-file with prototypes and #defines

25

26 //===Connects serial to ODrive and make ODrive object=== 27 SoftwareSerial odrive_serial(8, 9);

28 ODriveArduino odrive(odrive_serial);

29

30 //=======Declaring constants======= 31 byte DOUTS[2] = {DOUT1, DOUT2};

32 long int results[CHANNEL_COUNT];

33 HX711MULTI scales(CHANNEL_COUNT, DOUTS, CLK);

34 int gain = 20;

35 int velocity = 0;

36 int max_velocity = 2000;

37 int min_weight = 10;

B.3. CODE WHERE ONLY ONE HX711 IS USED 39 float weightDist;

40 float volt;

41 bool start = false;

42

43 //=======Setup is run once======= 44 void setup() {

45 odrive_serial.begin(115200); //Establish communication between ODrive and Arduino

46 Serial.begin(115200); //Comment out when not

connected to PC

47 Serial.flush(); //Make sure all data in

buffer is sent

48 tare(); //Tare load cells

49

50 odrive.SetVelocity(0, velocity); //Make sure velocity is 0

51

52 pinMode(LED_BUILTIN, OUTPUT); //Use led on pin 13 to see when Arduino is asleep

53 digitalWrite(LED_BUILTIN, HIGH); //Turn LED on to indicate Arduino is awake

54 }

55

56 //=======Loop that will run continuously======= 57 void loop() {

58

59 //Check if there is enough power in the LiPo batteries supplying the ODrive

60 checkVoltage(); //Call function checkVoltage()

61 if (volt < 28.5 && volt >= 28) { //When battery is getting close to empty:

62 batteryLow(); //Call function to

make board go slower 63 }

64 else if (volt < 28 && volt > 26) { //If battery is empty: 65 batteryEmpty(); //Come to a full stop

and put Arduino to sleep (safety having ">26" as a

66 } //single faulty

reading happens from time to time) 67

68 //Read the load cells

69 readLoadCells(); //Returns array, results[1] is back of board

APPENDIX B. CODE FOR THE ARDUINO 70

71 if (start == false) { //Replays

until minimum weight is reached

72 if (weightDist > min_weight) { //Check if weight requirement is met

73 delay(3000); //If it is

, wait 3 seconds so the person can get up fully on the longboard

74 readLoadCells(); //Read

load cells again

75 weight = results[1] * calibration_factor; //Weight read, will be the same as distribution when you get

on

76 start = true; //Set

start to true so rest of code can run

77 }

78 }

79

80 //In this loop the commands that runs the motor are called depending on weight distribution

81 if (start == true) { //If minimum weight is reached, this block of code runs

82 if (weightDist < 0.9 * weight) { //If leaning forward

83 leanForward(); //Call function

leanForward and accelerate

84 }

85 else if (weightDist > 1.1 * weight) { //If leaning backward

86 leanBackward(); //Call function

leanBackward and brake or reverse

87 }

88 else { //If standing

straight

89 standingUpright(); //Call function standingUpright

90 }

91 enoughWeight(); //Check if weight

still is enough, if not: stop 92 }

93 }

94