Utvärdering av testutrustning och metoder för kabelnötning

(1)

Cable abrasion equipment construction and methods evaluation

ANDREAS EDLUND

Master of Science Thesis Stockholm, Sweden 2007

(2)

(3)

Cable abrasion equipment construction and methods evaluation

Andreas Edlund

Master of Science Thesis MMK 2007:14 MDA282

KTH Industrial Engineering and Management

Machine Design

SE-100 44 STOCKHOLM

(4)

(5)

Examensarbete MMK 2007:14 MDA282

Utvärdering av testutrustning och metoder för kabelnötning

Andreas Edlund

Godkänt

2007-02-27

Examinator

Jan Wikander

Handledare

Mikael Hellgren

Uppdragsgivare

Scania CV AB

Kontaktperson

Dan Magnusson

Sammanfattning

När Scania ska välja elektromekaniska komponenter till sina fordon är det viktigt att de testas noggrant innan de tas i produktion. Detta görs för att försäkra sig om kvalitén för att minimera risken för fel. Eftersom antalet elektriska kablar i fordonen växer så blir utrymmet mindre och påfrestningarna på höljena ökar. Det här examensarbetet handlar om metoder för att testa kabelhöljen mot nötning. Det finns två metoder för detta standardiserade i ISO 6722. I examensarbetet har jag konstruerat en maskin enligt sandpappersmetoden och jämfört den med nålmetoden. Efter konstruktionen av maskinen har tester visat att den uppfyller alla krav för att kunna jämföras med nålmetoden. Slutsatsen av den jämförelsen är att det inte finns någon uppenbar korrelation mellan de olika metoderna. Det är också svårt att säga om resultatet går att relatera till riktig nötning i fordonen. Mitt slutliga intryck är att i förlängningen kan nålmetoden med vissa justeringar ersätta sandpappersmetoden.

(6)

(7)

Master of Science Thesis MMK 2007:14 MDA282

Cable abrasion equipment construction and methods evaluation

Andreas Edlund

Approved

2007-02-27

Examiner

Jan Wikander

Supervisor

Mikael Hellgren

Commissioner

Scania CV AB

Contact person

Dan Magnusson

Abstract

When Scania chooses electromechanical components for their vehicles they are tested carefully before taken into production. This is to assure the quality of the components and minimize the risk of failure. Since the number of cables in a vehicle is growing the strains at the covering is getting bigger. This thesis handles the methods for testing the cable covering against abrasion. There are two methods vstandardized in the ISO 6722. In this thesis I have constructed a machine according to the sand paper method and then a comparison with the scrape method has been made. After the construction I have found that the machine fulfills the demands that were set up. The conclusion of the comparison and evaluation of the two

methods shows that there are no obvious correlations between the different methods. There is also not possible to say if they are related to real abrasion in the vehicles. My impression of the different methods is that in a long term and with some adjustments the scrape abrasion test can replace the sand paper abrasion test.

(8)

(9)

Preface & Acknowledgments

As a student you should at the end of the Master of Science in engineering education perform a Master Thesis of about 20 points. The Master Thesis can be performed either at the institution or a company. This thesis have been performed at Scania CV AB in Södertälje.

The thesis is made at the department of machine design at KTH. I’d like to give a special thanks to my supervisor at Scania Dan Magnusson my supervisor at KTH Mikael Hellgren and my manager at Scania Joakim Gimholt. A thanks also to the people at RECT and RECU for the help you’ve given in different matters.

(10)

(11)

7. References ... 38 Appendix A Signal Diagram ... II Appendix B Mechanical Drawings...III Appendix C Circuit Diagram...IX Appendix D Port Schedule ... X Appendix E List of components ...XI Appendix F Program Code... XII Appendix G Table of abrasion results...XXI

(12)

(13)

1. Introduction

Scania CV AB¹ is a company with a global production of trucks, buses and marine engines.

With more than 30000 employees and production in both Europe and South America the sale was almost 60000 vehicles in 2005[1]. Though 96% of that sale is outside the borders of Sweden all the research and development at Scania is concentrated to Technical Center in Södertälje. My Master thesis “Construction and evaluation of methods for cable abrasion”

was performed at the Technical Center of the RECT² department.

The responsibilities of RECT are the following:

• Assure the quality of electro mechanical components that will be used in production.

• Installation and repairs of electronic components and systems in the test vehicles at the Scania laboratory.

• Contribute with knowledge in quality assurance and system architecture in pre developing projects concerning electrical components.

• Support the departments of construction in electro technical matters.

1.1 Background

When Scania is about to choose electromechanical components for their trucks and buses there are several factors affecting the choice. Some of them are price, quality and length of life. To ensure that the quality of the components is the same as Scania demands (and the supplier promises) all components are tested carefully before taken into production. This is done to minimize the risk of failure in the vehicles.

To avoid misunderstanding it’s important that the supplier and Scania speak the same language in terms of environmental conditions, loads, time etc when they communicate around a component and its related tests. To ensure this, standards are used. A standard is a document that is written so that testing and measurements, to mention some, are performed under equal circumstances. There are different sorts of standards, Scania for example, has its own standards but there are also national and international standards. The biggest and most well known organisation for standardisation is ISO³ and it’s an ISO standard this master thesis is based on.

ISO 6722 handles cables in road vehicles. The standard describes everything from how to measure the thickness of a cable to the colour codes of the covering. Since the cables are such a big and important part of the trucks it’s of big importance that the choice of a cable is made carefully, there are several factors that have to be taken in to consideration when choosing cables. Examples are connectivity, heat resistant, stiffness and resistance to abrasion. As the number of electronic components in the vehicles just is getting bigger and bigger the cables are taking more space and the strains at the cover are growing. In ISO 6722 one chapter handles the plastic cover resistance to abrasion and how it should be tested. There are two

1 Called Scania in the report.

2 Enviromental Testing and Electromechanics

3 the International Organization for Standardization

(14)

different methods for abrasion testing in ISO 6722, the scrape abrasion method and the sand paper method. At Scania there has been a wish of testing all the cables that are purchased with the methods specified in the ISO 6722. Beside the methods specified in ISO Scania also want to have the possibility to perform extended variations of the different methods. The commercial machines that exist on the market today don’t cover these needs and therefore in 2004 a bachelor thesis were written at RECT, covering the construction of a scrape abrasion testing machine. That machine has been used since then and the scrape abrasion method is the most common abrasion test in Europe. The sand paper method is common in USA and in Sweden there is only one existing machine today. ISO 6722 tells nothing about how relevant the tests are according to abrasion in the vehicles or how the two tests correlate, therefore Scania wants to design a sand paper abrasion testing machine to examine this. This master thesis covers the design of such a machine and the testing and evaluation of the two different methods.

1.2 Test methods

To get an overview over the ISO 6722 and the two test methods a summary of the standard and information about the two testing methods follows.

1.2.1 ISO 6722:2002

”Road vehicles – 60V and 600V single-core cables – Dimensions, test methods and requirements”

ISO 6722 covers electrical cables in road vehicles up to 60V DC and electrical cables in road vehicles 60-600V. These are two different classes and extra care needs to be taken when working with the higher class. The standard handles single-core cables or one core in a multi- core cable as long as it falls within the same parameters as the single-core. In ISO 6722 dimensions of both the cable and the covering and how they are measured are specified. It also specifies different test methods and what requirements the cables should fulfil. Different areas that there are test methods specified for are electrical, chemical, mechanical and even flame propagation. The abrasion testes are located under a chapter that covers resistance to abrasion. Both methods is only applicable for cables , cables with bigger area don’t need abrasion testing. Next section consists of more information about the two test methods.

6mm2

≤

1.2.1 Scrape abrasion test

A 1 m test sample is prepared (4). This sample is then placed under a force of (3).

A needle is then abrading in the longitudinal direction of the cable for a length of . Four measurements are made. A measurement consists of abrading until the needle gets in contact with the conductor. After each measurement the test sample is moved 100 mm and rotated 90° clockwise. The numbers of cycles are counted and the cable should fulfil a requirement that the supplier and customer have agreed to.

05N , 0 7±

1mm 5 , 15 ±

(15)

Figure 1. Scrape abrasion test

1.2.2 Sand paper method

In the sand paper method a 1 m test sample of the cable to be tested is prepared. A special sand paper with conductive stripes at every 75 mm and the test sample are mounted according to figure 2. The pivot arm, bracket and rod are applied with a pressure of on top of the cable. Above the cable an additional weight according to a special table depending on the thickness of the cable is applied.

05N , 0 63 ,

0 ±

The sand paper is then drawn with a speed of1500±75mm min. When the paper gets in contact with the cable the length of sand paper is measured and the test sample is moved 50 mm and rotated 90° clockwise. This procedure is repeated four times and the mean value of the four measurements is the test result. The requirements are depending on the dimension of the cable and they are presented in a table.

(16)

1. Rod

2. Additional mass 3. Pivoting arm 4. Test sample 5. Bracket

6. Tape supporting pin 7. Sand paper

Figure 2. Sand paper abrasion test

The supplier and the customer decide which method that is to be used and there are no recommendations in the standard of which method to use.

(17)

2. Method

The reason this master thesis is made, is that the two methods should be evaluated against each other. To do this there must be something to evaluate and that is where the construction part of the master thesis comes in. These two things together divide the thesis into two natural parts, construction and evaluation. The report is also divided in this way, where the first part is about the design and construction of the test equipment. This part also consists of an evaluation of the equipment to see if the results are good enough to be used in a comparison with the other machine.

The second part treats everything about the testing and evaluation of the different methods.

This part is a bit different. Every single test is presented with method, purpose and results. At the end of the chapter there is a conclusion where a discussion is held whether there is a correlation or not between the methods. If there is no correlation, is it because the tests represent different kinds of abrasion or maybe one of the methods can be discarded?

2.1 Pre study

A Master thesis should start with a pre study where the writer get an understanding what the Master thesis is really about and also examines if similar things have been performed before.

This is often made as a literature study. Naturally there is not so much information to get in this subject, therefore the literature studies mostly has been to read different standards and study the bachelor thesis about the scrape abrasion method that were made at the department two years ago. There has also been a study over the old machine to see if there are any solutions that can be used. In addition to these two things some abrasion testing machines for other purposes like fabrics, concrete and laminated sheets have been looked at to get an understanding how different solutions might look like. During the work there has been extensive reading of data sheets and manuals to learn about the different components. This is something that always needs to be done when working with microelectronics. If there is going to be a microcontroller in the machine the software is depending on which microcontroller to be used and that is also something that needs to be learned during the time.

(18)

(19)

3. Problem description

In chapter one and two there is a short overview what the goal and outlines of this master thesis are. This chapter contains a more detailed description of what is to be accomplished. As mentioned in the background the cables in a truck are exposed to wear from different components. This kind of wear is called abrasion and that is what’s to be tested. There are machines at the market today that fulfils the demands for an ISO-test. The problems with these machines are that they only fulfil the demands for an ISO-test. Scanias wishes are to perform several different variations of this test and today there are no such machines in the market. To get a better overview of the demands a requirement specification follows that takes into consideration the requirements of both Scania and ISO.

3.1 Requirement specification

In the specification I’ve chosen to list the ISO demands first and then list the Scania demands.

This is done so it will be easier to follow and follow up. In the beginning of chapter four there is a diagram to illustrate the requirements.

ISO

• Exert a force of 0,63±0,05N from pivoting arm, bracket and support rod.

• Sand paper abrasion tape speed 1500±75mm min.

• Additional mass 0,05 – 1,5 kg

• 150J garnet sandpaper tape with 10 mm conductive strips perpendicular to the edge of the sandpaper spaced a maximum of every 75 mm.

• Record sand paper length needed to expose the core of the test sample.

Scania

• Adjustable speed, 1500±50%mm min

• Adjustable cycles, pre determined length of paper to test.

• Possibility to use regular sand paper.

3.2 Problems to be solved?

What ISO does is to describe how the test should be performed according to speed, angles and masses and what type of sand paper tape that is to be used. What the power train should look like and how to attach the mechanics and similar things are not treated in the standard. These problems are in the hand of the constructor to solve and some of the fundamental problems are treated below.

Feeding

The first problem to solve is how to move the paper from A to B via the cable. This is probably the most important thing to find a well thought through solution to. If the feeding of the paper doesn’t work the testing can’t be performed. The feeding will be motorized and the

(20)

challenge is to combine the controlling of the motor with a design that can move the paper without sliding. What is the best way to design it in order to have a good solution both for the feeding and the control of paper speed?

The precision of the paper speed is important to get correct and it’s a rotating motion that will be transformed into a linear motion. Here a consideration of mechanical design and minimum and maximum speed of the motor must be done. Should the control circuit be a simpler circuit which is not so flexible or should it be a more advanced and flexible solution with a microcontroller?

Length measure

According to ISO the test will go on until the core of the cable is exposed. At that time the test stops and the length of the used paper is measured. The standard prescribes that a certain type of paper, with conductive stripes at the surface is to be used. The accuracy of these stripes is quite low since the stripes are placed with a distance of 75 mm in the paper.

To decide the length of the sand paper that has been used is not so hard if a feeding wheel solution is used. Only a count of the numbers of laps is necessary, this is under the condition that the radius of the feeding wheel isn’t changing.

Weight regulation

Concerning the weight regulation ISO has 5 different weights for different thicknesses and areas of the cables[2]. Here it must be taken under consideration whether this is enough for the Scania testing as well or if a continuous weight scale is needed.

Fixture

As the sand paper is abraded against the cable, the cable will have a tendency to slide forwards in the direction of the abrasion in the horizontal plane. This put some requirements of the fixture of the cable since the abrasion should be against the same point at the cable all the time. Hence the construction can not be too complicated to work with at the same time it should hold a good grip of the cable.

End of test

Both in the case of ISO abrasion and in the case of a manual settled abrasion there need to be a good way to know when to stop and to actually stop it without bend on the accuracy in the measurement. For the ISO test it needs to be a separate sensor to detect while it for the manual test should be enough with the system for keeping constant speed and stop after a predetermined number of cycles.

Evaluation

Everything above has been problems around the construction. The second part of the master thesis is about the testing. How do I know that the machine works properly and fulfils the demands of ISO 6722? Apart from testing every component and part of the system during the construction the complete machine needs to be tested the same way several times to show that it repeats the same result over and over again. This is to show that when using the machine a test only need to be performed once and that the machine itself not is a source of error. Earlier testing with the scrape abrasion machine has shown that an important factor is the temperature. The plastic covering behaves differently if the surrounding temperature differs

(21)

more than 5-10°C. In the requirements for the scrape abrasion method it’s specified to keep the temperature at . For the sand paper abrasion test this requirement don’t exist but it is good to keep in mind that the temperature might affect the result. 23±1°C

If the machine works as intended but there is no correlation with other methods at least it can be used for comparative tests between different cables. When the machine is approved the second part of the master thesis starts. This part is to compare the scrape abrasion test with the sand paper abrasion test to see if there is any correlation between the two methods and as far as possible see if there are any similarities with the abrasion in the trucks. If there is a correlation between the two methods it might show that one method is better than the other to use and by only test with that method a result with the other method can be estimated. This is advantageous if the test time between them differs a lot.

(22)

(23)

4. Cable machine

After defining the task in chapter one and then framing the problems bound to the task in chapter 3 the work has at this point come to where the final design is laid and the construction of the machine starts. The first thing is that a few solution concepts to each of the bigger problems are weighted against each other and from them the best fitted for the task is chosen.

When this is done the construction of the machine is described in terms of mechanics, electronics and programming.

4.1 Conceptual Design

In the conceptual design some different solutions are considered and pros and cons are discussed and from there a conceptual solution is chosen. What’s important to understand is that it’s a concept for a solution not the exact solution that is presented. Conceptual design begins with an illustration of the requirement specification to clearify that every aspect of it is handled in the construction of the machine.

Sand Paper Method

ISO Manual

Adjustable length

Adjustable speed

Regular sand paper

05N , 0 63 ,

0 ±

75 min

1500± ^mm 5kg , 1 05 ,

0 −

°

±

° 2 29

Record length

Figure 3. Requirement specification

4.1.1 Speed/distance

There are several solutions for keeping a constant speed. At an early point the decision fell at an electric motor for the feeding of the paper and from that starting point some possible solutions have been studied.

1. Optical reading

In this solution the paper is rolled up at the feeding wheel. Next to the roll an optical sensor is mounted that reads the thickness of the roll and gives feedback to the motor. The length measurements are more difficult to calculate as well when the radius of the roll differs.

(24)

Optical sensor

a

Figure 4. Alternative 1 & 2 for cable suspension Advantages:

+ Known starting speed + Collects the paper Disadvantages:

− The accuracy in the readings

− Bulky design with the external sensor

− Hard to control since the motor speed needs to be reduced when the roll gets thicker

− Measurement of length

2. Tachometer at suspension pin under paper

Similar to 1) but instead of the optical reader there is a tachometer connected to a) that has a feedback to the motor.

Advantages:

+ Precise reading of the sand paper length that has passed the cable + Collects the paper

Disadvantages:

− Unknown starting speed Æ Adjustment time

− More difficult to control due to increasing thickness of the roll

3. Tachometer attached to motor

The sand paper is fed with the motor and a rotating load b) after the roll the paper is piled up at the table.

b

Figure 5. Alternative 3 for cable suspension Advantages:

+ Precise controlling since the tachometer is attached to the motor axis

(25)

+ Always one layer of paper which gives a ”plug and play” function + Known starting speed

Disadvantages:

− The paper is not collected.

− Skid between motor and load due to bad friction.

Since every paper should be used only once the advantage of rolling up the paper is minimal.

Therefore the first method is discarded and from the other two the best thing would be a combination. To limit the amount of work I’ve chosen the last one because it’s easier to implement and gives a more accurate result.

4.1.2 Adjustable weight

The adjustable weight can be solved in different ways. The two main ways are either a manual solution with different weights that are applied depending on what kind of cable that is to be tested. Solution number two is an automatic solution with some kind of motor and sensor. Advantages with the weights are that it’s easy to control that the weight is exact and that it is a self controlling system during the abrasion. It’s also an easier solution to implement. Disadvantages with this solution are that it demands loose parts and that it is discrete. It also might have a tendency to slip of the cable during testing.

The automatic solution is more flexible since it is continuous for all values between zero and max. It’s also easier to handle for the user since the only thing needed is to set a weight in the display. When it comes to disadvantages the biggest is that there is no way that the user himself can control that the correct weight is applied. The user has to trust the machine to make the correct adjustments. The automatic solution is anyhow a nicer and more mechatronical solution. The decision has fallen on a combination of the two solutions.

Primary solution is the automatic and as a backup, weights can be used.

1. Linear movement

A linear actuator placed at the point of rotation pushes a weight on top of the pivot arm.

Advantages:

+ Easy calculations, only torque around one point.

+ Cheap, no external sensors are needed.

+ Self regulating during the abrasion.

Disadvantages:

− The motor is only concentrated around the centre of gravity when the pivot arm is horisontal(since the movement of the arm is so small this shouldn’t be a problem).

− It’s hard to get enough span of weight since the pivot arm is not long enough.

2. Circular movement

The actuator presses up the pivot arm to the right of the centre of rotation, by this solution the cable located to the left is pressed down. With an angular sensor and knowledge of the force from the actuator the pressure at the cable can be calculated.

(26)

Advantages:

+ No movable mass. The span of weight is therefore limited only by the actuator.

Disadvantages:

− Extra sensor Æ higher price.

− Advanced calculations with sinus and cosine, unnecessary stress of the μC.

3. Direct working actuator

Above the cable and the pivot arm an actuator affects a load cell that is applied at top of the bracket.

Advantages:

+ Precise feedback with the load cell.

+ Big span of weight Disadvantages:

− Price.

− No self regulation.

Because of the decision to have a combination of both an automatic and a manual weight regulation the third alternative is best suited for the task. It’s also important that the adjustment of the force is precise and the solution with a load cell is best suited for this.

4.1.3 Cable attachment

To keep the attachment of the cable user friendly there are some demands. The basic demand is to fix the cable during testing. To keep the machine robust the attachment shouldn’t consist of loose parts that fall of during testing. The third demand that has to be fulfilled is that it must work for different cable dimensions. Preferably all attachment should be without any external tools to make the machine independent.

There are some different solutions that should be taken into consideration when choosing solution. One option is to use crocodile clips. Easy to use and no loose part makes them good for the task. What can be a little risky is their ability to hold a steady grip of the cable at bigger loads. Another solution would be similar but instead of having crocodile clips the cable would be attached with screws in a right angle from the cable, the same principal as a terminal block. This solution is not perfect in the user friendly view of it but instead it will hold a steady grip of the cable. The last is an automatic solution, but this I’ve discarded because of the complicated design. Even if none of the first two is perfect I’ve chosen the second one to prioritize the grip in front of user friendliness.

4.1.4 System control

To get hold of all the controlling of the system and to have a functional user interface electronics are needed. The challenge with the interface is to make it simple and clear but still having a good functionality. This is to minimize the risk of user related errors and as far as possible give the machine a pleasing look. A possible solution is some kind of two- or four

(27)

row LCD⁴ and a control for choosing row and value. A button to confirm the choice will also be used. To make the interface work and still have a good functionality a microcontroller is needed. My choice of microcontroller will be a PIC and there are two reasons for that. It’s possible to use C as the programming language and it’s the kind of microcontroller that is being used at the department hence giving access to necessary knowledge when needed.

4.1.5 Motor

The starting point of the construction is the motor. This is the only component that is decided before the construction has started. The motor that is to be used is a brushless 28V fan motor.

To control the motor a PWM⁵ or an analogue signal is needed. Both the analogue and the PWM works from ground to supply voltage. To control the motor with a PWM it needs to be fed at a frequency of 2 kHz ±0,5 kHz. Since it is a fan motor the thought is too only demount necessary parts and use the motor for cooling of the system as well. The motor is shown in figure 6.

Figure 6.The motor of the system.

4.2 Construction

4.2.1 System analysis and microcontroller

An analysis of what needs to be done is made before the building of the machine starts. Even if it might take some time to think through the design it is well worth it to get things right from the beginning. The analysis is most important in the choice of microcontroller. If a μC with too few pins or one that is missing a certain function is chosen it could be fatal to the project plan.

4 Liquid Crystal Display

5 Pulse Width Modulation

(28)

Reviewing the requirement specification and the conceptual design it seems that it will take quite a lot of pins just to fill up the connections from the peripheral components. A list of the components is showed in appendix E. In the list is also presented what different functions the components need to be controlled. These functions can be directly translated to modules in the μC. The choice of μC has fallen on PIC18F458. The PIC18 is a RISC⁶ CPU with 40 pins and it has all the functions that are needed. Once the choice of μC is made a structure needs to be set up to make sure that the ports are used the best way. This is to fit all connections from the peripherals and to make the programming as smooth as possible. In appendix D a table of the port structure is shown.

4.2.2 Power supply

The power supply is a switched embedded unit, which minimizes the risk of someone to supply the electronics with wrong voltage. The first criterion I’ve looked at is the maximum desired voltage. The biggest consumer in the system is the motor that is fitted to 28V but will work up to 35V. Studies of the data sheets and diagrams of the motor shows that 24 V is enough to meet the demands of the system. The motor is also the biggest current consumer in the system and it can consume as much as 14 A depending on load and speed. Maximum speed for the machine will be approximately 3000 rpm which is a little more than half of the maximum speed. At the motor the current not is proportional to the voltage the biggest consumption at full speed will be 5A.

Maximum power consumption.

W A

V 5 120

24 ⋅ = (1)

Since the rest of the components mostly are 5V IC:s with currents at mA level the motor is by far the biggest power consumer. With these assumptions the chosen aggregate is at 240W and 10A. This is to ensure that there will be no dips caused by bad power supply.

In the system, only the motor and the holding solenoid demands 24V. Most of the components need 5V and therefore needs to be lowed. Since IC:s are very sensitive to stabile feeding voltage to work properly the choice has fallen on fixed voltage regulators at 5V respectively 15V because these are more stabile than an own voltage divider.

4.2.3 Mechanics

In Conceptual design different partial solutions were selected and in the analysis the electronics were combined to a system The next thing to do is to combine the mechanical parts to a complete system in order to dimension it. The complete mechanical system is shown in figure 7.

6 Reduced instruction set computer

(29)

Figure 7. Mechanical system

The dimensions of the different parts in the system are depending on the dimension of the sand paper to be used. The ISO paper is only 30 mm wide and due to the demands from Scania that regular sand paper should be compatible with the system as well, the dimensioning is to be fitted to the regular paper where the smallest rolls have a width of 50 mm.

In order to dimension gears and the radius of the feeding wheel I began dimensioning the passive parts as the roll that will press the sand paper to the feeding wheel and the roll to which the sand paper rolls will be attached. When this was done a mechanical model over the system was made where the forces are presented so it is possible to calculate the maximum torque needed to run the system. This is crucial for the choice of gears and dimensions of the wheels for the tachometer and feeding. First the dimensioning of the passive parts was done.

If nothing else is said, aluminium with a density of 2.7^kg_dm³is used. All inner frictions from rolling bearings are neglected since they are very small compared to other forces.

Passive parts

Since the chosen solution consists of two rolls of which only one is connected to the motor the other need to be easy rolling yet stable to keep the paper in place. Due to this the machine must be flexible to different thicknesses of sand paper. This leaves two opportunities, either move the feeding wheel or move the stabilizing wheel.

The stabilizing wheel solution consists of an angled axis to which a cylinder is attached via two rolling bearings in each end of the roll.

(30)

Figure 8. Stabilizing wheel

The shape of the axis comes from the demand of stability. The wheel is movable in vertical direction and to prevent the wheel bumping from vibrations when running a holding solenoid will keep the wheel in place. The space is necessary to fit in the solenoid above the axis of the feeding wheel. The stabilizing wheel is bigger than the feeding wheel in order to get the paper in a horizontal line during the feeding and the size is set to ∅24 mm with ∅19 mm roller bearings passed in to each end. If the two bearings and the cylinder are approximated as solid bodies the moment of inertia for the cylinder will be:

2 6 1

2

1 3,15 10

012 , 0

0438 , 0

2 J kgm

m r

kg r m

J m ⇒ = ⋅ ⁻

⎭⎬

⎫

⎩⎨

⎧

=

⇒ =

= ⋅ (2)

In the extension the moment of inertia can be used to calculate the force that needs to be applied to spin the roll.

( )

r Jt F r

F t M

J

M ⇒ = ⋅

⎭⎬

⎫

⎩⎨

⎧ ⎟ = ⋅

⎠

⎜ ⎞

⎝⎛ =

⇒

⋅

=α α ω , ω (3)

The maximum sand paper speed,v_p, in the specification is 2250 mmmin and with a width of 24 mm at the roll the angular velocity is given by.

(

^p ^d

)

^rad^s

v 3,125

60

2 ⇒ =

⋅

= ⋅ ω

ω (4)

To minimize the risk of under dimensioning we assume a rise time that is very short t = 0.1s.

This gives us all the values needed to calculate the force.

N F

r t

J r J t

F_fw _fw ³

6

10 21 , 8 012

, 0

1 , 0

125 , 3

10 15 , 3

−

⋅

=

⇒

⎪⎪

⎭

⎪⎪

⎬

⎫

⎪⎪

⎩

⎪⎪

⎨

⎧

=

⋅

=

⋅ ⇒

= ω ω

(5)

(31)

The same calculations are made for the sand paper roll. The biggest roll that can be used is a 50 m roll with a diameter of 22 cm.

N F

r t J r Jt

F_sp _sp 0,3

11 , 0

1 , 0

341 , 0

10 68 ,

9 ³

=

⇒

⎪⎪

⎭

⎪⎪

⎬

⎫

⎪⎪

⎩

⎪⎪

⎨

⎧

=

⋅

=

⋅ ⇒

=

−

ω

ω (6)

Since the sand paper roll just hangs on a stiff axis there is a friction between the roll and the axis. This is a force that needs to be taken into consideration. The coefficient of friction between the plastic inner of the roll is assumed a higher value than the real.

N mg F

mg

F_f^sp _f^sp 9,42

6 , 0

7 ,

15 ⇒ =

⎭⎬

⎫

⎩⎨

⎧

=

⇒ =

⋅

= μ μ (7)

The width of the axis that the paper hangs on has a width of 6,9 mm, knowing this makes it possible to calculate a force that can be added to the force needed to spin the roll.

N F

F

F_f^sp _x^sp _x^sp 0,29

2 0069 , 11 0 , 2 0

0069 ,

0 ⎟⇒ =

⎠

⎜ ⎞

⎝

⎛ +

⋅

=

⋅ (8)

Add (7) and (8) and the total contribution to the force from the sand paper is received.

N F

F F F

F

F _f

sp x sp sp

x sp f

f 0,59

29 , 0

3 , 0

3

3 ⇒ =

⎪⎭

⎪⎬

⎫

⎪⎩

⎪⎨

⎧

=

⇒ = +

= (9)

Another part that is needed for the calculation of forces is the force from the pivot arm, bracket and additional weight. The pivot arm and the bracket should give a total force of and the maximum applied weight is 1,5 kg. The bracket is calculated to weigh 50g and this is a force that is applied right above the cable. The pivot arm is instead attached to the machine and has a momentum around one point. To translate this into a force at the cable a rigid body in horizontal position is calculated at. This gives a force that is

05N , 0 63 ,

0 ±

m g F_pivotarm = ⋅

2 (10)

at the cable. The last part, that is passive is, the steering pin that is placed to get the right angle up to the abrasion point. The paper won’t have a big pressure against the pin and even though it’s the rough side of the paper that is in contact with the pin an assumption is made that this force is so small it can be neglected.

Forces and Torques

When all the forces from the passive parts are calculated a mechanical model of the system can be set up. In the model there are only the forces that are essential for the calculations, the rest are removed to keep the picture clean.

(32)

Mechanical model of the system:

F1 Weight

3

Ff 1

Ff Cable

Stabilizing wheel

N1 Sand paper roll

Supporting pin

2

Ff

Fm

Ffw Steering pin

rm

Feeding wheel

Figure 9.Mechanical model

The torque from the motor is a combination of all counter forces and the radius of the axis.

m m

m F r

M = ⋅

0 : − ₁ − ₂ − ₃ − ≥

→ F_m F_f F_f F_f F_fw (11)

63 , 0 ) 81 , 9 5 , 1 ( 63 , 0 63

,

1 0

1 = F + =mg+ = ⋅ +

N

The coefficient of friction for aluminium and the different sides of the sand paper is unknown but an assumption is made that the friction on the rough side is significantly bigger than the paper side therefore the paper side is neglected. Tests that I’ve made gives a coefficient of friction between paper and cable about 0,7. To make sure that the forces aren’t under dimensioned a slightly higher coefficient than the real is assumed.

N N F

N

F_f _f 13,81

9 , 0

345 , 15

1 1

1

1 ⇒ =

⎭⎬

⎫

⎩⎨

⎧

=

⇒ =

⋅

= μ μ (12)

All the forces are now known and they can be added to get the total force the motor needs to produce.

N F

F F

negligible F

F F

F F F

F _m

fw f f f

fw f

f f

m 14,41

10 21 , 8

59 , 0

81 , 13

3 3

2 1

3 2

1 ⇒ =

⎪⎪

⎭

⎪⎪

⎬

⎫

⎪⎪

⎩

⎪⎪

⎨

⎧

⋅

=

⇒

−

=

−

(13)

14,41 N is the total force that the feeding wheel must manage to pull the paper at. This force can be translated into a torque depending of the radius of the feeding wheel. This is treated in the next piece.

(33)

Feeding axis

From the motor via the gearing is the feeding axis. It’s a solid axis with different radius. A tachometer is mounted on the feeding axis. The plate has 32 holes for keeping the right speed.

A closer description of that will be given in the “Programming” chapter. At the end of the axis is the part called feeding wheel. With a feeding wheel that has half the radius of the stabilizing wheel a torque that is calculated in equation (14) is given.

(14)

r Nm r F

F M

m m m

m

m 0,17

012 , 0

41 ,

14 =

⎭⎬

⎫

⎩⎨

⎧

=

⇒ =

⋅

=

This torque is slightly less than the motor can produce without gearing. This means that the critical factor will not be the torque but the speed. The limitation is not the maximum motor speed because it will always be possible to shift down. The motor stops if it drops below ~600 rpm therefore a lower limit to 1000 rpm is set. The down shift needed is the quota between the rotational speed of the motor and the rotational speed of the feeding wheel.

fw m

v

i= v (15)

Rotational speed of the feeding wheel is given by sand paper speed and radius of feeding wheel.

(

^r ^mm

)

^v ^rpm

v v _fw _fw

fw p

fw ⇒ ∅ = ⋅2 =37,68 ⇒ =19,9

=∅ π (16)

This gives an approximately gear shift ratio of:

9 50 , 1000 ≈19

=

i (17)

With a desired sand paper speed of 2250, the speed of the motor need to be:

rpm i

v_m 2985

68 , 37

max = ⋅2250 = (18)

The calculations of the gear ratio above are only approximations but they give a good direction when choosing what kind of gearing that will be used and what values to choose to get close to the value above.

Gears

Based on the values above, the gear chosen is a transmission belt gear. This is a quiet, easy to handle gearing. The ratio of transmission in the system is rather high, and therefore it’s necessary to shift it down in several steps. The wheels are standard and the down shift will be in 3 steps, 72:15, 72:15, 72:30. Total ratio of the gearing is

(34)

84 , 30 51 72 15 72 15

72⋅ ⋅ =

=

i (19)

With the gear ratio calculated, all the values needed for programming and dimensioning of the electronics are available.

Covering

The machinery and all of the electronics that not need to be handled by the user are embedded in a box. The box is a system of aluminium profiles. Front, sides, top and bottom are then made of aluminium plates of different thickness. The bottom and front are thicker since they need to carry the weight of the mounted components.

4.2.4 Electronics

All peripherals in the machine are connected to a circuit board. The circuit board is a combination of surface- and hole mounted components, which makes it a compact solution.

It’s from the circuit board everything are controlled and if there is a problem here it affects the whole system. Electronic components are sensitive from outer interference and to shield the electronics on the board everything has been embedded in a casted aluminium box. On the board all voltage inputs and signals has been decoupled with capacitors to ground to filter the noise. To further prevent disturbances all copper in the board that not are solders are grounded and as a further precaution the aluminium box is grounded as well.

The input voltage to the board is 24V but as mentioned under Power Supply most of the IC:s need 5V. There are 3 regulators on the board, one for 15V and two for 5V. One 5V goes to the linear actuators and one goes to the rest of the board and the reason for this is that testing has shown that the load cell signal is interfered by the actuators.

The system is provided with ICSP⁷. The ICSP makes it possible to program the system at place. Programming is made by open the box and connect the black cable inside to an external ICD⁸.

Connectors

Since the circuit board is embedded there is a need to connect to the peripherals on the outside. This can be made by wiring all the cables through a hole but it would make it an inflexible solution if something needs to be demounted. Instead I’ve chosen to have external connectors to the box. However this brings the risk of someone connecting the wrong way. To prevent the user from making that mistake the ambition has been to make every connector unique. This has been done with the exception of some 4 mm standard plugs that are used at two different connections. But with a clear marking wrong connection should be avoided. The 24V input is protected by a diode and a fuse to prevent the system to burn if someone connects ground and voltage the wrong way.

Signal Amplification

The pins of the PIC18 are CMOS⁹ and TTL¹⁰, which means that they will have logic levels between 0- 5V. This means that the signals in and out from the μC are not always in the

7 In Circuit Serial Programming

8 In Circuit Debugger

(35)

correct levels from the peripherals. To deal with this some kind of amplification of the signals are needed. There are two signals in the system that need to be gained, the control signal to the motor and the signal from the load cell.

The control signal to the motor goes via an OP that is built up as a non-inverted amplifier where the size of the two resistors decides the gain. The signal is 5V and the output signal should be 15V therefore a gain of 3 is desirable. The gain can be estimated as shown in equation (20).

73 , 11 2

30

2 1 2

1 ⇒ =

⎭⎬

⎫

⎩⎨

⎧

Ω

= Ω

⇒ =

= G

k R

R

G R (20)

The load cell gives a differential signal and with a needed gain of 500 a normal connection to an OP isn’t enough. In this case an instrumentation amplifier does the work. An instrumentation amplifier consists of two OP:s. With a differential input it transforms a negative and a positive input to one positive output. The gain is set with an external resistor,R_G, and are decided by equation (21).

RG

G kΩ

+

= 50

1 (21)

Peripherals

All the components outside of the aluminium box that are controlled by the box are referred to as peripherals. These include the holding solenoid, the stop sensor, the user interface, the tachometer, the load cell, the motor and the linear actuator. The holding solenoid is there to lock the stabilizer roll once the paper is in place. It is driven by 24 V but should be controlled by the μC which can provide 5V, this is solved by connecting the solenoid via a transistor.

The load cell and linear actuator together forms the force regulation system. This is a system that will not be used in this version of the machine. This is because the load cell has lost its calibration during testing and it will take too long time to get a new one. Nevertheless the circuit board is prepared for connecting the components. It is also prepared in the program code for implementing this function.

A standard 2 row LCD¹¹, a press able pulse encoder and a switch is the user interface. The switch is for powering the machine. After that the user gets all the information on the display and makes the choices by turning the encoder and confirms by pressing it.

4.2.5 Programming

The base of the programming has been to create an easy to use interface. But the easier it’s to use the more complex the programming becomes. This has set some limitations to the original idea. The following chapter is a general description of the programming, the complete code can be found in appendix F.

9 Complementary Metal Oxide Semiconductor

10 Transistor-Transistor Logic

11 Liquid Crystal Display

(36)

User Interface

The interface is built up by the step-by-step principle. This means that the user has two options for every choice or the possibility to set a value and once the user has confirmed the choice there is no opportunity to go back one step. There are some places in the program where the user is asked if the program should be aborted. The structure of the menu system is shown in figure 10.

Figure 10. Menu system flow chart

Three times during the menu cycle the user has the opportunity to abort. The three times are after the cable attachment, before starting the test and during the test. If choosing to abort the user get the opportunity to either go back to where he were or to restart the test procedure from the start. The option to abort isn’t implemented in every step due to programming technicality, more of that under structure.

Code structure

Even though the menu system is linear the code isn’t. It would be easier to make the code linear but with the choice of making the user interface easy and only use one button the user would be locked and if something goes wrong the only way to fix it would be to restart the machine. Instead of doing that the code is based on interrupts.

This means that when a certain event takes place, for example the user presses a button, an interrupt method is called and an action connected to the event that just occurred is

Abort

Start

Attach

Shut down

Length Speed Mass

ISO Manual

Mass

Start

New Test?

Stop

Method Cable

Test done

(37)

performed. With this solution the same method can be used for abortion no matter where in the code the user is at the moment. Since the interface to the user consists of only one button the actions must be more software controlled and that demands more from the programming because the μC needs to know where in the code the user is all the time.

That is done by setting flags and variables in order to perform what the user expects. This is one reason why there are only three spots in the program where the user can abort. For every extra event it’s more complicated to get hold of all the flags and the code have a tendency to swell. It also is a bigger risk for bugs, and the most important task is the functionality of the machine not to have a perfect user interface. The other reason is that when a program jumps between methods there will be uncompleted tasks, this results in stacks that are growing with time. An μC have limited memory space and after a while the memory is filled up and bugs will appear due to lack of memory.

After the user has set all the values and they’ve been saved into the memory the user starts the procedure. When it starts the PWM-register is set to the desired value. The calculations are shown under algorithms. Since it’s difficult to get the speed exactly correct with the calculations only, a control loop is connected to the tachometer at the axis. Every time the counter gives a signal an interrupt is generated at the ECCP¹² port. The microcontroller then measures the time between two signals and adjusts the speed up or down depending on what’s needed. This is measured 32 times pro lap and therefore adjusts the speed quickly.

At breakthrough or at desired length the PWM is set to zero and the motor stops immediately.

The result is then calculated by the numbers of time the CCP¹³ has gotten an interrupt. The result is displayed to the user.

Algorithms Time

Maximum time of μC-timer in capture mode.

ms t

bit prescaler

MHz f

bit prescaler

t f _C

osc

C 105

65536 16

8 20 16

4 *

1 _max

max ⇒ =

⎪⎭

⎪⎬

⎫

⎪⎩

⎪⎨

⎧

=

∗ ⇒

= _μ

μ (22)

This means that the maximum time between two measures can’t exceed this time unless an extra variable is set every time the timer overflows. The time is important to know when the tachometer is to be made so the numbers of holes will be correct. To keep it more simple and accurate the aim is too not exceed 105 ms. For calculation of the holes the first thing to know is the angular speed of the feeding wheel, . Divided by 60 the laps pro second are received. v_r

rads v_r v

π π

⋅

∅

= ⋅ 60

2 (23)

When the time is known the distance between two holes, , is easily calculated. s_r rad

v t

s_r = ⋅ _r (24)

12 Enhanced Capture/Compare/PWM module

13 Capture/Compare/PWM module

(38)

One lap is 2π and divided by the number of holes that fits in one revolution is held. s_r

sr

n= 2π (25)

A combination of equation (23), (24) and (25) gives the minimum number of holes in the encoder. To be on the safe side a time that is shorter than the micro controller time is used.

15 , 30 12

1 , 0

750 min 2

60

2 ² ⇒ =

⎪⎪

⎭

⎪⎪⎬

⎫

⎪⎪

⎩

⎪⎪⎨

⎧

=

∅

=

⋅ ⇒

⋅

∅

= ⋅ n

mm s t v mm v

n t

π

π (26)

The minimum number of holes that are to be needed is 31. The encoder is made with 32 holes as mentioned in the mechanical chapter. This is done since it’s easier to do an even number of holes. These numbers can be used to produce an algorithm for the reference time and an algorithm for keeping the speed to match that time.

t v n

mm v

t_ref n _ref 70,65

32

60 12 ⇒ =

⎭⎬

⎫

⎩⎨

⎧

=

⇒ ∅

⋅

= ∅ π

(27) When divided with the maximum allowed μC time, a percentage share is received. Multiply the share with 65536 and the reference value that is to be compared with the value at the ECCP module is received.

ECCP v

t t

C

ref 44096366

65536

max ⋅ ⇒ =

μ

(28)

Equation (27) and (28) could be merged into one equation but the compiler can’t handle that.

Therefore in the code they are divided into separate calculations but the extra time doesn’t matter since it’s not in a time critical phase of the program. For every capture the actual time is compared to the reference time and the speed is adjusted.

Speed

The speed is decided by a PWM, but it’s not the frequency of the PWM that is decisive, it’s the average voltage. Even though the motor can be driven by the frequency of a signal the PIC CCP-module works in much higher frequencies than the signal frequency of the motor. The motor can therefore not discern the pulses instead it interpret the pulses as an analogue voltage.

The PWM-module is a 10-bit module. With a gain of ~3 over the OP-amplifier the maximum signal to the motor is measured to 15,17V. Readings of the status signal the maximum rotational speed of the motor will be 2900 rpm at that voltage. With the shift down of 51,84 the maximum sand paper speed will be ∼2100 mm/min. Now the quota between wanted speed and the value to the PWM can be determined.

Utvärdering av testutrustning och metoder för kabelnötning

Cable abrasion equipment construction and methods evaluation

Cable abrasion equipment construction and methods evaluation

Andreas Edlund

Sammanfattning

Abstract

Preface & Acknowledgments

Table of Contents

1. Introduction

2. Method

3. Problem description

4. Cable machine

( )

(

)

(

)