The Facility for Applied Manufacturing Enterprise

Student Research Projects

Students: Martijn Christenhusz
Patrick Pol

Project: Coffee Buddy

Supervisor: Dr. J.S. Burns
College of Engineering
San Diego State University
5500 Campanile Drive
San Diego CA 92105
USA
Phone: (619) 594-6067
Fax: (619) 594-6005
Email: jburns@mail.sdsu.edu

Thesis advisor: Ir. J.de Heer
Hogeschool Arnhem Nijmegen
Ruitenberglaan 26
6828 CC Arnhem
The Netherlands
Phone: 026-3658181
Fax: 026-3645066
Email: dhr@tfarn.nl

SUMMARY

An injection molding machine (IMM) creates thermoplastic coffee cup covers The produced covers vary in quality: they can be bent when their cooling time is too short or contain airbubbles when the solid plastic pellets were not completely dry. Short-shots can occur when the mold is not completely filled, or the sprue breaks off and gets stuck in the mold.

Some kind of assembly line has to be designed that can check the covers on shape and clarity. The sprue has to be broken off the cover, and the covers have to be weighed and packaged. Weighing data has to be stored in a file. The whole assembly line has to be educational for mechanical engineering students.

First the covers have to be aligned, because they fall randomly aligned out of the IMM.
This will be done mechanically with a metal gutter.

A big S110r six axis robot, which has its own controller, will be used to manipulate the covers. One major problem is that the robot is mounted to the floor at approximately 10 meters from the IMM, and neither machines can be moved. Transportation from the IMM to the robot will be done rolling the covers between two C-shaped rails.

Weight will be measured with a load cell with a resolution of at least 0.1 gram.
Clarity and roundness will be checked spinning the cover with the robot, while sensors check the edge of the cover for roundness and the surface of the cover for clarity.

The process will be controlled by two devices; a PLC and a computer running LabVIEW 4.0 software. LabVIEW will be used to save data into computer files, display process parameters on a graphical user interface, and approve or reject the covers by measuring their quality. The PLC controls the actual assembly line (switches and actuators), and communicates with the robot. LabVIEW and the PLC will communicate over an RS-232 serial interface by sending ASCII characters.

The packing of the covers is not a part of the project yet. For now, the covers will be placed in squares of four covers on the packaging table, and packed manually. In the future, a packing machine can be added to the process line.

A main task of LabVIEW is data acquisition (DAQ). LabVIEW measures several parameters, like signals from the quality check station, a load sensor for measuring weight, and temperatures on the IMM, Temperatures will be measured with thermocouples and RTDs. To be able to do data acquisition, LabVIEW needs DAQ hardware boards and NI-DAQ software. A National Instruments AT-MIO-16E-10 data acquisition board will be plugged in to a PC. The NI-DAQ 5.0 software allows LabVIEW to communicate with its hardware board.
TABLE OF CONTENTS
SUMMARY 2PREFACE 61 INTRODUCTION 72 DEFINITION STUDY 82.1 PROJECT 82.2 PURPOSE 82.3 PLANNING 82.4 THE INJECTION MOLDING MACHINE (IMM) 92.4.1 Principle 92.4.2 The BOY 15/5 Injection Molding Machine 102.5 MATERIALS 122.6 COFFEE CUP COVERS 132.6.1 Non-fatal error 142.6.2 Fatal errors 142.7 PROCESS DIAGRAMS 152.8 SYSTEM DEMANDS 173 FUNCTIONAL DESIGN 193.1 PROCESS STRUCTURE 193.2 ALIGN COVERS 193.3 REMOVING BENT COVERS 203.4 SPRUE BREAKING 203.5 QUALITY CHECK 213.5.1 Roundness check 213.5.2 Clarity check 213.6 WEIGHING DEVICE 223.7 PACKAGING 223.8 REJECTED COVERS 233.9 ENVIRONMENT 233.10 MAN-MACHINE DIALOGS 243.11 SYSTEM DEMANDS 244 TECHNICAL DESIGN 254.1 SENSORS 254.1.1 IMM sensors 254.2 ALIGNING THE COVERS 264.3 REMOVING BENT COVERS 274.4 SPRUE BREAKING 304.5 BUFFER FOR COVERS 324.6 WEIGHING DEVICE 334.7 QUALITY CHECK 344.7.1 The roundness check 344.7.2 Clarity check 364.8 PACKAGING THE COVERS 384.9 THE GMF S110R ROBOT 384.10 OVERCOMING THE DISTANCE BETWEEN ROBOT AND IMM 394.11 THE TOTAL PROCESS LINE 414.12 PROCESS CONTROL 414.12.1 Sensors of the process line 444.12.2 The Programmable Logic Controller (PLC) 454.12.3 LabVIEW 464.12.4 Emergency stop wiring 464.12.5 The PLC-LabVIEW communications channel 475 LabVIEW Error!5.1 DEFINITION STUDY 475.1.1 Project 475.1.2 Planning 475.1.3 What is LabVIEW? 475.1.4 The assembly line. 505.1.5 LabVIEW tasks 505.1.6 Requirements 515.1.7 The Input / Output 525.2 FUNCTIONAL DESIGN 525.2.1 The initialization screen user interface 535.2.2 The main program user interface 545.2.3 Program initialization 565.2.4 The main program 595.2.5 Program termination 605.2.6 Communication with the PLC 635.3 TECHNICAL DESIGN 635.3.1 The initialization screen user interface 635.3.2 The main program user interface 645.3.3 Initialization 645.3.4 Main program 655.3.5 Program termination 685.3.6 Communication with the PLC 685.3.7 The computer setup 696 Data acquisition 706.1 MEASURING TEMPERATURES FROM THE IMM 706.1.2 RTD temperature elements 706.1.3 Thermocouple measurement 736.2 ANALOG SIGNALS 766.3 DIGITAL SIGNALS 766.4 THE HARDWARE BOARDS 776.5 REQUIREMENTS FOR THE HARDWARE BOARDS 78Authors list 80APPENDIX I GLOSSARY 82APPENDIX II TIME SCHEDULE 85APPENDIX III IMM SPECIFICATIONS 86APPENDIX IV WEIGH TEST RESULTS 87APPENDIX V DRAWINGS MODULES 88APPENDIX VI SENSOR SPECIFICATIONS 88APPENDIX VII ROBOT SPECIFICATIONS 94APPENDIX X AT-MIO-16E-10 SPECIFICATIONS 100

PREFACE

When we were searching for an appropriate project for our final study in the Netherlands, we were happy to learn that there was a project for us available at San Diego State University in San Diego, U.S.A . After checking our map of the United Sates and intensive communication with this university, we accepted the assignment under the direction of dr. J. Burns and dr. G. Bailey at the Facility for Applied Manufacturing Enterprise (FAME).

The resources and technical equipment that this facility at the Mechanical Engineering department has are amazing for a future engineer, and are a source for enthusiasm and inspiration.

This report is our final report for our electrical engineering study at the HTS in Arnhem, the Netherlands, and it discusses the project we worked on during our four months stay in the United States. It discusses parts of electrical engineering, automation and mechanical engineering, sometimes referred to as mechatronics. This report is also supposed to be a firm base for other San Diego State students to continue the project we started.

We would like to thank everybody who helped us making our four months stay as pleasant as possible, or helped us with the project in any way. We especially would like to thank the Mechanical Engineering department of the San Diego State University for making this thesis possible for us, Ir. J.H. de Heer for his support from overseas, and dr. J. Burns and dr. G. Bailey for their outstanding support, inside and outside the university.

San Diego, United States, may 1997

Martijn Christenhusz
Patrick Pol

1 INTRODUCTION

Injection Molding is a method of forming plastics and powdered metals into various shapes. It has many advantages over other material processing techniques: flexibility, very good dimensional accuracy, excellent surface finish, and the capability to produce highly complex shapes.

The faculty for Automated Manufacturing Enterprise at the San Diego State University have an Injection Molding Machine (IMM). The covers produced by this machine are semi-finished and their quality is not constant. The sprue still needs to be broken off and the shapes of the covers can vary.

For educational purposes, a process has to be designed to create a process which is able to inspect a particular part, a coffee cup cover, directly after they are produced. Therefor this process is designed and uses many different processing techniques to create a high educational value. For example: this project uses LabVIEW, a PLC, a six-axis robot, sensors, switches and different types of transportation.

The size of the project is too big to finish in 4 months, so the main goal is to make a firm foundation for other graduate students. The PLC and the robot programming part should be done by two SDSU senior students for a project. This project will be continued the next few months by these two students.
2 DEFINITION STUDY

2.1 PROJECT

An Injection Molding Machine, available in the laboratory, creates thermoplastic coffeecup covers, or coffee buddies at a rate of three to four per minute. The coffee covers can be used as a cover for a coffee cup. Due to these covers, warm liquids stays longer warm and with a logo of the school, it can be used as promotion material. The produced covers vary in quality: they can be bent when the cooling time is too short or contain airbubbles when the solid plastic pellets are not completely dry. Short-shots can occur when the mold is not completely filled, or the sprue breaks off and gets stuck in the mold. If this happens the machine is not able to produce a new cover anymore, until the sprue is manually removed.

This multi disciplinary project is started by us to make a foundation for other students, so they can finish the project. The whole project is estimated to last at least one year. Our project consists out of the development of a complete system that is able to inspect high impact polystyrene coffeecup covers on a number of defects like shape and clarity. The process line must have a quality inspection for the covers and the data collected along the process line must be stored by a process control system. It has also to weigh and package the covers. The resources for this project are limited. If possible, resources already available in the laboratory must be used. Examples of availability are a conveyor belt, LabVIEW (virtual instrumentation soft- and hardware), a PLC, a six axis robot, and actuators.

2.2 PURPOSE

The purpose of this mechatronic project is to create a cover inspection and assembly line which is educational for Mechanical Engineering students of SDSU and can deliver covers that could be commercially sold. For the educational part of the project, many different processing techniques available in the laboratory must be used. This is in general terms a restricting factor for the solutions of this project.

2.3 PLANNING

The project size is too big to finish it in time. Our goal is to make a base for other students so they can finish this project. According the planning, it is not possible to finish the technical study in time without assistance. This assistance is provided by two American Mechanical Engineering students from SDSU. Their assistance starts in the technical design phase, after the decisions are made which tools and methods there are going to be used. For the time schedule of this project, see appendix II.

2.4 THE INJECTION MOLDING MACHINE (IMM)

Injection Molding is a method of forming plastics or powdered metals into various shapes. The process involves the rapid pressure filling of a mold cavity with a fluid material, followed by the solidification of the material into a product.

2.4.1 Principle

The injection molding process of thermoplastics can be divided into three stages. The first stage is the plastification stage; in this stage, the plastic pellets are transported from the feed hopper to the screw. The heating zone melts and homogenizes the material in the hydraulic screw/barrel system. The screw, however, is allowed to retract, to make room for the molten material in a reservoir at the cylinder head, between the screw tip and the nozzle valve. The nozzle heating is meant for keeping the melted plastic pellets in the viscous liquid state. (See figure 1).

The second stage is the injection stage. At the injection stage, the screw is used as a ram (piston) for the rapid transfer of the molten material from the reservoir to the cavity of the closed mold. Since the mold is kept at a temperature below the solidification temperature of the material, it is essential to inject the molten material rapidly to assure complete filling of the cavity. A high holding or packaging pressure (10,000-30,000 psi or 600-2000 atm) is normally exerted, to partially compensate for the thermal contraction (shrinkage) of the material upon cooling. This cooling of the material is a big limiting time constant in injection molding, because of the low thermal conductivity of polymers.

Figure 1 Injection Molding principle

After the cooling time, the mold opens and the solid cover is ejected by the ejector pins in the mold. This is the third stage and called the ejection stage, and shown in figure 2.
Figure 2 Mold part ejection principle

Many thermoplastic resins require thorough drying prior to molding, to avoid the formation of voids or a degradation of the material at the molding temperatures. Therefore prior to the feed hopper, a hopper dryer is used to get the moisture out of the plastic pellets.

2.4.2 The BOY 15/5 Injection Molding Machine

The "BOY 15/5" is an old, made in 1974, relatively small Injection Molding Machine capable of producing about three to four covers per minute, depending of the amount of plastic injected and the cooling time of the cover. The barrel heating is divided in three zones: heat zone 1, heat zone 2 and the nozzle heating. The nozzle heating is regulated by power adjustment between 0 and 100°C, (see figure 3). The temperature of the two heat zones are separately adjustable with analog regulators between 0 and 450°C. The accuracy of these regulators is not very good anymore, but that is not that important for this process. The heat zone temperatures are measured by J-type thermocouples which are made of a Fe-Const. Constantan is an alloy of copper and nickel. The most used constantan variant for thermocouples contains 57% copper and 43% nickel. The normal temperature range for this kind of thermocouple is between -190°C and +760°C.

The heat zones are used for the melting of the plastic pellets, while the nozzle just keeps the plastic at temperature. The zones are melting the plastic pellets while they are transported through the tube by the screw.

This IMM can operate in three modes:

Stepper mode: the machine steps through the production cycle and after each step the operator has to activate the next step by pushing a button.
Semi-automatic mode: the IMM goes through the whole cycle creating one cover, then it must wait for the operator to press a button to start the next cycle.
Automatic mode: the machine produce covers without interaction of an operator as long as there is enough basic material and the machine does not malfunction.

Figure 3 Heating zones of the BOY 15 IMM

The three heat zone parameters are adjustable at the control panel of the IMM. For semi-automatic and automatic mode, two more parameters are required. The cooling time of the cover in the mold, (the time that the mold stays closed), so the cover can cool down. The other parameter is the injection time, which is the speed of the injection. After opening the mold, the covers are rejected and fall randomly out the machine. Out of sixty covers, 56% fell up side down. The outlet of the IMM is placed 0.65 meter above the ground.

Most movements systems on the IMM are hydraulics; the mold is directly closed hydraulically and the material is also injected by a hydraulic ram. The machine is equipped with a mechanically-hydraulically acting safety valve which is called the sliding guard. On opening the sliding guard, the oil feed to the clamping mechanism is shut off so that no further movements can take place. The clamping mechanism can only travel if the sliding guard is closed.

The manual of the IMM is limited and does not contain much information about the electrical system of the machine. The useful specifications of the IMM can be found in appendix III. The IMM will probably be replaced by a more modern and faster machine.

2.5 MATERIALS

All thermoplastics are, in principle, suitable for injection molding, but since fast flow rates are needed, grades with good fluidity (high melt index) are normally preferable. In our case, polymers are used.

A polymer is a very long molecule and many of these molecules together form the plastic. The plastic is built from lots of unorganized strings of polymer molecules. The polymer we use is an amorf (without shape) polymer, polystyrene. Crystallizing polystyrene is not suitable for injection molding. There are different stages for amorf polymers, this depends on the glass transition temperature (T_g). Below T_g the polymer behaves like a solid material and the polymerstring can not move with regard to each other. If the temperature gets higher than T_g, the polymerstring can move a little with regard to each other, so the material is less stiff and more rubber-like. When the temperature is raised more, the polymerstring can move relatively easy with regard to each other and the plastic becomes viscous liquid-like. Polymers never become thin fluids, because of their molecular weight. The changes from solid-state to viscous liquid-state are reversible.

It is essential to realize that this solid-like to liquid-like transition, commonly but sometimes incorrectly is called 'melting'. Meltingpoints or meltingranges are not dependable on time. The T_g on the contrary is time dependable, i.e. quick heating or cooling results in other T_g's than slow heating or cooling.

Polystyrene is one of the major thermoplastics. It is just like many other plastics commercially available in a wide variety of grades and variations. There are many different kinds of homopolymers, like "conventional," "normal," "regular," "standard," "general-purpose" (GP), or "crystal polystyrene". Impact resistance grades are usually referred to as "high impact" (HI) or "rubber-toughened polystyrene". The used material for the manufacturing of the cover is high impact polystyrene, see figure 4. The homopolymer is of the noncrystallizing type and below its glass transition temperature T_g(100°C), it is very stiff but brittle and highly transparent glass-like material.

Figure 4 molecular structure of polystyrene (PS)

Polystyrene tends to have a rather low intrinsic resistance to weathering, and UV radiation causes yellowing and further embrittlement. The fire resistance of polystyrene is not very good. Many organic fluids, such as aromatic and chlorinated hydrocarbons as well as food oils and fats, are either solvents or cause undesirable changes to the material.

The most common processes of polystyrene are: Injection molding such as low-cost disposable service wear, cabinets, toys etc., Extrusion of finished products such as profiles, pipes as well as sheets which are subsequently thermoformed into small or large objects.

2.6 COFFEE CUP COVERS

Figure 5 Image of cover with sprue

The covers are made of high impact polystyrene, and weigh approximately 26 grams. The exact weight depends on the density and the used pressure during the fabrication of the cover. The weigh test results can be found in appendix IV. It is also possible to make covers of dyed plastic pellets which results in covers with another color.

Figure 6 Dimensions of the cover in centimeters

2.6.1 Non-fatal error

There are different kinds of cover errors which can occur during fabrication regarding the shape of the cover: one non-fatal error and three fatal errors. The non-fatal error for the cover, occurs when the sprue gets stuck in the entrance to the cavity. When this happens the machine must be stopped as soon as possible, since the machine is not able anymore to produce new covers. The sprue must be removed out of the entrance of the cavity. The cover without sprue can continue its way in the process line. This is a fatal process error but the part can be still on specification. This error currently occurs in approximately 3% to 5% of the produced covers, but a new Injection Molding Machine has a far lower malfunction rate. One of the reasons for sprues getting stuck is the damage at the mold cavity entrance, which is not that smooth anymore because of abuse of the mold.

2.6.2 Fatal errors

The following cover errors are fatal:

If there is not enough material in the nozzle to fill the whole mold, the produced cover is oval instead of a desired round cover. This kind of error is hard to detect, especially when there is a little short shot. The process must be able to detect short shots that are obvious by eyesight, allowing short shots up to approximately 1mm. A short shot is when the diameter X of the cover at some point 1mm smaller is (diameter Y) than the normal round cover. If the machine is kept properly adjusted and the nozzle does not leak, short shots do not occur frequently (<= 1%). According the operator of the IMM, the set amount of the plastic drifts a little in time. This can be noticed after 60 to 70 covers are produced.

Figure 7 Short shot from a cover

The plastic must be completely dry before it can be used for the process. If there is any moisture left on the plastic pellets, it forms steam in the viscous liquid-like plastic. Once in the mold the steam forms bubbles in the plastic during the solidification and these bubbles are visible when transparent polystyrene is used. Limits for clarity are hard to specify, and can not be given yet. This error hardly occurs (< 1%) and only if the plastic pellets are not thoroughly dry.
If the cooling time, after the plastic is injected into the mold, is to short, the plastic is not completely solid. If the cover is rejected from the mold and falls out of the IMM, the soft plastic cover can deform. The cooling time depends on different parameters. The mold can use water as coolant, but it is not regulated by the machine. The operator has to adjust the water flow manually. The mold gets warmer if the machine is producing many covers.

Deformation can also occur during the ejection stage, if the cover gets stuck to the mold. These deformed covers must be removed out of the process as soon as possible. Because of their different shape are they sometimes hard to handle.

2.7 PROCESS DIAGRAMS

Figure 8 shows a simple schematic overview of what the process is expected to do. Covers can be rejected by the process for different reasons; they may be bent, contain air bubbles or have short shots. The covers that are approved should be packaged in packages of 4.

Figure 8 Process schematic overview

The processing block can be divided in smaller en more specific modules. This is shown in figure 9 on the next page.

Align cover: This module must align the covers which fall randomly out of the IMM, so the covers reach the filter module in the same position.
Is cover flat?: The filter selects the bent covers out to the rejection box, because bent covers can cause serious problems further along the process line.
Has cover a sprue? If the cover has not a sprue, the next module does not have to try to break the sprue off. Covers with a sprue lose their sprue.
Break sprue off: The covers with a sprue must be positioned so their sprue can be broken off.
Is the cover round? All the remaining covers are checked for their roundness. If the cover is not round, the cover is rejected and falls into the rejection box.
Is the cover clear? All the remaining covers are tested for their clarity. The covers which fails the test, are rejected and end in the rejection box.
Weigh cover and store data: The approved covers are weighed and their data must be stored in a data log file.
Place covers in package of 4: The covers are placed in clusters of 4. In the future a packaging device can be attached.
Remove cover from process flow: The rejected covers are collected in a box. These covers can be recycled, so there is no material lost.

Figure 9 Flow diagram of total process

2.8 SYSTEM DEMANDS

Because of the educational purpose, the process installation is not allowed to be higher than 1.5 to 2 meters. Otherwise, students would not be able to follow the whole process anymore. The process has to be designed in small modules, which allow moving the process or parts of it. It must be possible to add in the future new modules to the process. It is not allowed to move the IMM from its original position. The operator must be able to control the process from one central place, nearby the process. The operator must be able to follow the covers in the process line. All data of the process must reach the operator, so he or she is able to make the right decisions. The covers must be weighed during the processing for statistic use only. All the data must be stored in two different files, one file where all events during the process are stored, like rejection of a cover. Another file is used for the storage of malfunction messages.

Of course, the safety aspect is also important. The IMM already has his own safety requirements, but the new to be designed system also must have its safety measures and equipment. It is not allowed to go to an unpredictable state when the operating system malfunctions. The total system has to meet these safety requirements. Covers must not get stuck in any process module. If this not can be guaranteed, some method of checking on these malfunctions has to be built into the system. The operator has to be notified on these malfunctions.

The system needs a modular design because of the possible future expansion of the process line. A possible expansion of the process line is the label unit. With this unit it is possible to print text or label every single approved cover. This must be placed between the weighing device and the packer.

Another change is the IMM. In the future the IMM will probably be replaced by a more modern and faster IMM.

3 FUNCTIONAL DESIGN

3.1 PROCESS STRUCTURE

After making the flow diagram in chapter 2.7, the functional designs of the different processing tasks can now be made. The production process includes the IMM, the transport between the modules, sorting and inspection of the cover. The process structure shows that the process is divided in different groups, shown in figure 10. The groups are handling, quality check and weighing.

Figure 10 Process structure

The handling of the covers can be divided in five new groups. Those groups are; aligning, removing bent covers, sprue breaking, packaging and rejected covers.The next paragraphs show the I.P.O's (Input Process Output) of the modules.

3.2 ALIGN COVERS

A test has proven that the covers fall randomly out of the machine. Therefore, some kind of method has to be designed to align the covers, so that they all can be processed in the same way. A mechanical sorting-method is preferable, because it has a low malfunction-rate, is inexpensive to maintain, and requires no power. This module must be able to align and orientates all the covers, regardless the quality of the covers. The covers have to be presented to the filtering stage.

The maximum number of covers to handle are 4 covers per minute. The covers need to be presented to the filter module, with an alignment accuracy of 2 mm. During the handling the covers are not allowed to get more additional damage on the surface.

Figure 11 I.P.O of align cover module

3.3 REMOVING BENT COVERS

Bad shaped covers, like bent ones must be taken out of the process flow as soon as possible for several reasons:
Bent covers are much harder to handle and they cause more interruptions and malfunctions in the process than on specification covers. It takes also extra effort and time to break the sprue off and weigh them, finally they are rejected anyway

A filter, see figure 12, has to be designed to separate the normal covers from the bad covers. The normal shaped covers pass the filter and the bent, or covers with a short-shot (too small) not. Two major errors must be detected separately, so a process controller can give the operator accurate information. A cover is not allowed to get stuck for more than 3 seconds. If a cover is rejected from the process, the operator must be notified within 1 second after rejection, and a filter rejection counter must display this information. The covers which are filtered out must be collected in a separate box, all other covers are allowed to continue their way.
Figure 12 I.P.O of filter

3.4 SPRUE BREAKING

The covers which have passed the filter come to the sprue breaker module, see figure 13. This module must be able to detect whether or not the cover has a sprue. If there is a sprue attached to the cover, the sprue must be removed. Covers without a sprue are transferred to the next module. The covers are still aligned to one position so the sprue can be broken off. The sprues which are broken off are not collected in a box, because it is unpredictable where the sprues land when they are broken off. Some kind of fence can be used to force the srprues in a predictable direction. The sprue breaking cycle may take 4 seconds. After the breaking process the covers continue their way sprueless.

Figure 13 I.P.O of sprue breaker

3.5 QUALITY CHECK

This module must be able to detect the covers which are not completely round or not completely transparent. The quality check module can be divided in two sub modules, one module for each check. The quality check module must be able to detect the covers which are not complete transparent, due to airbubbles. The other sub module checks the roundness of the cover.

3.5.1 Roundness check

Covers which have a small short shot can pass the filter, so they need to be checked accurately on roundness, see figure 14. The module checks the outside of the cover, in case of an error , the cover must be rejected.

If a cover with a short shot reaches the roundness check, the check must be able to detect a short shot , resulting in a size difference of 1mm over the X/Y axis. The roundness data must be shown in a graphic. The roundness check may take no longer than 4 seconds. If a cover fails the test, the cover is rejected from the process flow. The rejected cover fall into a rejected cover box. When this happens the operator must be notified and a counter shows how many covers have failed the roundness check.
Figure 14 I.P.O of roundness check

3.5.2 Clarity check

The clarity check must be an option, since dyed covers can not be checked on clarity.

If a cover with airbubbles reaches the clarity check, they have to be taken out of the process flow. The check must be able to detect airbubbles bigger than 1 mm, and the clarity check must be done within 8 seconds. If a cover fails the test, the cover is rejected from the process flow. The rejected cover falls into a rejected cover box. When this happens the operator must be notified and there must be counted how many covers have failed the clarity check. The data of the clarity check must be displayed to the operator. Figure 15 shows a input-process-output diagram of the clarity check.

Figure 15 I.P.O of clarity check

3.6 WEIGHING DEVICE

The weighing device, see figure 16, must be able to weigh the cover and it must be done at a resolution of tenths of grams. The purpose of the weighing device is for statistic use only, not for quality check of the covers. The weighing of a cover must be done within a second. The data for the weighing must be stored in a weighing data file.

Figure 16 I.P.O of weighing device

3.7 PACKAGING

If the covers successfully pass the quality check and the weighing, they finally reach the packer, see figure 17. The packer must pack the covers in sets of 4 and place the covers in a square. The number of packages made must be counted and made visual to the operator.

Figure 17 I.P.O of packer

3.8 REJECTED COVERS

The rejected covers have to fall into a box under the modules in the process line, see figure 18. The boxes must be emptied by the operator. The covers can be ground to raw material and recycled for new production.
When a box is filled with rejected covers they can be thrown into a granulate machine which make little pellets of the covers. So there is no material lost, except for the sprues, but they can also be recycled if wanted. The pellets must be dried first before they can be reused.

Figure 18 I.P.O of rejected cover box

3.9 ENVIRONMENT

It is important to design in this phase an environment suitable for humans, who are going to work with the machines. All the machinery must be adjusted to man and not the other way around. The following factors are important: human factor, work environment and operator qualities.

Human factor:
There should be no extremely heavy labor for the operator, only the removing of a jammed cover in the process line, emptying the recycle boxes and feeding of the IMM with plastic pellets. The rest of the time the operator can monitor the process from his desk. An operator must be able to control the process flow and can stop the process when it is necessary.

Work environment:
The room where the controls are placed is the same room where the IMM is installed. The room is a laboratory and filled with machinery. The room has no windows and has only one door which leads to the outside, and daylight can not penetrate the laboratory. The temperature is 16°C to 25°C, depending on the season, because there is no air conditioner available in the laboratory. The IMM does not make a lot noise when it is producing parts. The process line will not be working the whole day, maximal a few hours per day. Therefore the operators does not need a special chair or a specially decorated room.

Operator qualities:
The technical knowledge the operator must have to operate the assembly line should not be high. The interaction with the process must be easy to understand and must tell the operator what to do. For this work one operator is enough but for safety measurements a second operator is recommended.
3.10 MAN-MACHINE DIALOGS

The operator must, during the process, be able to see what the different parameters are. An important parameter is the temperature of the IMM heatzones. The operator must be able to read the temperatures on a display, placed at the operators control desk. The operator can not influence the temperature of the IMM heating zones from behind his desk. Perhaps this feature can be added in the future for another student project. The operator must be able to switch the clarity check on or off. Another important parameter is the number of produced covers. It must be displayed at the operators desk, together with the number of approved covers and the total rejection ratio. This ratio shows the total percentage of rejected covers in the whole process line. For the complete user interface, see paragraph 5.2.1.

3.11 SYSTEM DEMANDS

The most important system demand is the safety aspect of the whole process. The robot can hurt people very easily. The robot must be surrounded by a fence. On the process line there must be an emergency button to stop the robot. The other modules are not dangerous, because there are no moving parts which are dangerous for people. The operator must be able to switch off all the modules. These safety precautions measurements must prevent any injury of the operator or other people working near the process line. The operator must be able to stop the process at any time at any module. This is the emergency stop. The emergency stop must stop the whole process within 1 second.

Covers are not allowed to get stuck anywhere in the process, since this block the whole process flow. If this can not be guaranteed, the system has to check if the covers got stuck. If a cover got stuck the operator has to be notified and the operator has to solve the problem manually. A way to check the process flow is to time different sections. If a cover enters a module, a timer must start. Within a certain amount of time the cover must reach the end of the section. When the timer expires the set time, the cover must got stuck somewhere. The actual time can not be given at the moment, since the actual technical design for the process is not ready yet.

The controlled system shut down, must shut the process line down in a controlled way. At first the IMM must be stopped, so it can not produce new covers. The covers go down the process line until the are packaged. At that time the whole process line is empty and the process line can be shut down. In case of a power failure the process must be set into a safety mode, so when the power comes back, the system must wait until the operator resets and starts the system.

4 TECHNICAL DESIGN

4.1 SENSORS

Before choosing the right sensor for each application in the assembly line, certain specifications about the application must be found out. For every single sensor application, the same questions must be asked. Because of the big variety of sensors, like size and performance, it is important to select the right sensor. The following questions were asked for every single sensor application.

Sensing distance
Supply voltage
output type/switching requirements
Load requirements
Target material
Target dimensions
Target finish
Target movement
Mounting requirements
Response time
Electrical conditions
Mutual interference
Ambient operating temperature
Environment
Waterproof
Color detection

Most of the factors are the same for all applications. However, all factors are important, and lead eventually to the right choice for the application. With these questions the amount of usable sensors in the Omron catalog were reduced to two or three types per application. After checking the prices and comparing their performances, the sensor was chosen. These steps were taken for every single sensor application.

4.1.1 IMM sensors

The IMM has different switches, which can be used as sensors to see if the cover falls out of the mold. The switch on the mold and on the outlet of the IMM can be uses as a detector. The first switch (numbered b17 in the IMM manual) is normally open and is used by the machine to see if the mold is open or closed. If the mold is closed, the switch is closed and a relay (d27) is activated. The plastic is injected into the mold and the plastic cools down for 10 to 15 seconds. After cooling down, the mold opens, also opening switch b17. The cover falls out of the mold, into the outlet of the IMM. This outlet is guarded by switch b10. If the cover falls through the outlet, the switch is triggered.

The signals of these two switches are mainly used to indicate if the IMM malfunctioned. Two major malfunctions can occur an must be solved manually:

If a cover with sprue gets stuck in the mold, switch b10 is not set after the mold has opened.
It can happen that the sprue breaks when the mold opens; the cover falls through the outlet, but the sprue is still in the mold cavity. The mold closes again and the IMM tries to fill the mold, but the viscous liquid plastic can not enter the mold and the plastic will be injected outside the mold. When the mold opens, no cover falls through the outlet.

To detect these two errors, a timer is started when the mold opening switch (B17) is triggered. If the outlet switch (B10 and indicating if a cover fell out of the mold) is not triggered within 3 seconds, the operator must be notified. If the outlet switch was triggered, the timer stops and must gives a reset.

The outlet switch can also be used to keep track of the covers; every time that the switch is triggered indicates a produced cover.

4.2 ALIGNING THE COVERS

Figure 19 Aligning covers with a gutter

Alignment can be done with some kind of gutter, where the covers slide through and all get aligned with the sprue facing up. A picture of this gutter is shown in figure 19. This method requires no power at all and has no moving covers.

To select a material for the gutter, some tests had to be done. The tested materials were available in the laboratory. In order to get an indication of the smoothness, a cover was placed on a sheet of material. The sheet was placed under an angle and the angle was measured with an electronic leveler. If the covers were placed on the sheet and they slided down, the test was repeated. The test with one kind of material ended after 5 successfully slides under the same angle. The results of this test are displayed in table 1.

Material 15° 16° 17° 18° 19°

Polished Wood x

Formica laminate x

Aluminum x

Stainless steel x

Table 1 results of the sliding test

The suggested construction material is stainless steel. Formica laminate is also a usable material, but it requires a construction of wood or metal to support the thin laminate sheets. Stainless steel needs no further maintenance, like the polished wood and Formica laminate.

Two cardboard model were made to do some initial tests on the design. After one model, some changes were made to improve the results. Testing the second model, the results were satisfying, but are expected to be even better with the final metal version, which has a much smoother surface than cardboard.

A Pro-Engineer drawing of the gutter can be found in appendix V.

4.3 REMOVING BENT COVERS

Covers which are not flat, which means more than 3mm bent from the original form, need to be taken out of the process flow as soon as possible with some kind of filter. The second part of the filter takes care of the covers which are too small. There is no drawing of the turner in appendix V, because of its complex form, this must be drawn in a program like Pro Engineer. If this filter is placed after the gutter (see paragraph 4.2), it can be assumed that all the covers are already aligned when they enter the filter.

One filter make-up is shown in figures 20a and 20b. In this solution, the filter consists out of two parts, checking the filter on shape and size separately. The shape filter must be placed before the size filter, because bent covers do not fit in the C-shaped rails. Covers which get stuck before the shape filter, can be removed with some kind of trap door, placed under the shape filter.

Figure 20a Shape filter

The size filter only needs to take out the covers that are really too small, so they can not give any trouble further down the process line. The cover is moved from a horizontal position to a vertical position, guided by a C-rail. The covers which are too small can not be supported by the C-shaped rail, and fall out of the rail. The critical off specification covers are detected further on at the quality check.

Figure 20b Size filter

Another method of shape/size detection is using a software package for recognizing a shaded image of the covers projected on a Charge Coupled Device (CCD). This method is displayed in figure 21. The cover is placed upon the lens of a CCD camera. A light shines upon the cover and a light receiver transforms the signal to a gray tone. The receiver is built out of many cells and are available in different sizes. Every single cell is able to store a gray tone depending on the amount of light they were exposed to. The 'picture' of the new cover must be compared with a stored 'picture' of a cover without errors. An algorithm is used to detect the errors, and depending on the result of the calculation the cover can be rejected. In this case the decision to approve or reject the cover is made by computer software.

Figure 21 CCD image recognition

Though the second solution more accurate is than the first solution, the final filter is built like the first solution. The CCD has good options for data acquisition and educational purposes, but the device and software to do such an image detection are not available in the laboratory. The money to buy a big CCD camera with frame grabber and processing capabilities is not available for this project, so the easier and less expensive solution, will be used in this project.

If a cover get stuck in the filter module, it must happen before the shape filter. The sensing if a cover got stuck in the filter is done by a sensor. For this application the sensor E3S-LS from manufacturer Omron is chosen. The sensor has a selectable dark-on / light-on operation. The light-on operation is useful to detect if the sensor is defect or if the wire is broken somewhere, also known as fail save. The E3S-LS is a photoelectric sensor with wide visibility, which makes use of a method better known as diffuse reflective. The sensor uses the target surface as reflector. If the cover remains in the beam, the output signal of the remains low. When the sensor output is low for more than 3 seconds, an actuator connected to the trap door underneath the filter must be activated. The data sheets of this sensor can be found in appendix VI.

The turner module needs another sensor type. The place where the cover can fall out of the rails can differ, so the sensor must have a range from at least 1.0 to 15 cm. The method of detection is retroreflective with a photoelectric sensor. The sensor must be placed under the turner, the transmitter/receiver housing on one side and the reflector on the other side. The most suitable sensor in the Omron catalog with these specifications is sensor E3S-R. If a cover falls out the rail, the cover is detected when it appears between the emitter/receiver housing and the reflector. As result the photoelectric sensor switch is opened, it falls also through the beam of the sensor. The sensor is in light-on mode so the output of the sensor gets low for a short time en when the cover has left the beam the output becomes high again. When the output gets low, the filter rejection counter must be increased with one. The data sheets of this sensor can be found in appendix VI.

A second sensor is placed at the end of the turner. The function of this sensor is to keep track of the covers and is also used for timing to see if a cover got stuck somewhere in the process line. If the mold is opened a timer is started, and if the cover rolls normally through the process line the cover must reach the sensor within 10 seconds. If the cover does not trigger this sensor, an alarm message must be sent to the operator. If the cover is rejected in the filter or turner module the timer must be stopped. The sensor starts a second timer which only can be stopped by another sensor placed half way the transport rail. The sensor specifications of this sensor site are the same as the second sensor in the turner module. For both applications the E2K-C sensor selected out the Omron catalog most suitable.

The E2K-C is very suitable capacitive proximity sensor. The normally closed option is here also used to detect whether or not the sensor is broken. If the cover approaches the sensor, the capacitive proximity sensor switch opens and the output signal of the sensor becomes low. The cover rolls further and the sensor signal becomes high again (+24 V). The data sheets and the decision list can be found in appendix VI.

4.4 SPRUE BREAKING

This module must be able to detect whether the cover has a sprue or not. If there is a sprue attached to a cover, the sprue must be removed. The sprue is massive plastic and has no special breakpoint. Breaking off the sprue can be accomplished in different ways. Two of the given solutions involve the use of a big Fanuc S-110R six-axis robot, which is available for the assembly line in the laboratory. Further information on the S-110R robot can be found in paragraph 4.9 and appendix VII.

The cover with a sprue rolls guided in two C-shaped extrusion rails, placed under a angle, to the sprue breaker. Since the covers are already aligned, the sprues are all pointed to one side. A metal bar is placed next to the rail. The sprue rolls against the metal bar and the cover stops. If the cover lays still, the robot breaks off the sprue, just by pushing it down. The metal bar does not stop the cover anymore and the cover rolls further without a sprue to the next module. The C-shaped rails in this section need to be reinforced because of the momentum on the rail during the push action.

This solution is actually a variant of solution 1. Only here the sprue is broken off by a metal bar, from which one side is connected to a wire and the wire is connected to a motor. When the cover is stopped by the sprueblocker, the metal bar is above the sprue. The motor is turned on and rolls up the wire. This forces the metal bar down and during its way down it breaks off the sprue. The metal bar is an easy and cheap mechanical way to be sure that the covers with sprue are stopped, and the sprue lands always in the same area. Unlike the first method, this method does not require the use of a complex robot. However, using the robot in this process is highly educational.

The cover with sprue rolls down a rail into a pre shaped form. The function of the pre shaped form is to be sure that the cover always stops at the exact same position. The robot has to pick up the cover at the sprue and moves the cover to another location. When the cover is fixed, the robot must break the sprue while holding the sprue. The fixed cover can now slide back into the process stream to the next module. The sprue is held in the robot hand, after the sprue has broken off, so they can for example be dropped in a box. The big disadvantage of this solution is the need for a special robot hand. One hand is needed to pick up the covers at the sprue and one hand for the quality check, for picking up the covers at the top.

First the robot breaks as many sprues as necessary to fill the whole buffer. When the buffer is filled with covers, the robot is allowed to move to the end of the buffer for the weighing of the covers and the quality check. The blocking pin releases the covers one by one to the test station. The robot continues testing the covers until four covers are approved and packed together in one package.
This means some covers may remain in the buffer, which is not a problem. The robot only needs to know how many sprues to break during the next cycle. This method creates also more time for packaging, because it gets the four covers delivered in a shorter time, after which it has time until the next four covers arrive.

Figure 22 Sprue breaking in rail, by robot.

The first solution is the one, which is going to be used in the process. The advantage of this breaking process, is that there is no need for a special robot hand. The metal bar is an easy mechanical method to be sure that the covers with sprue are stopped at the spruebreaker station. This method, displayed in figure 22, is inexpensive and reliable, sketches of this module can be found in appendix V.

For the detection of the cover the sensor E3C-DS1 from manufacturer Omron is chosen, which is a Fiber optic sensor. The sensor is placed behind the spruebreaker see figure 22. The normally closed option is here also used to detect if the sensor or wire is broken (fail save). If the cover rolls through the sensor beam, the sensor switch opens. If the cover stays in the sensor beam, the output signal of the remains low. After the robot has removed the sprue, the cover rolls further and the sensor signal becomes high again (+24 V). The advantage of this solution is that there are a couple of these sensors available in the laboratory. In appendix VI, the data sheets of this sensor can be found.

4.5 BUFFER FOR COVERS

Moving the robot to and from the sprue breaker and the quality check station costs precious time.

Creating a buffer for the covers before the quality check station allows the robot to work in series, instead of switching between his tasks all the time. This buffer is shown in figure 23. The C-shaped rails before the sprue breaking station can work as the actual buffer.

Figure 23 Cover buffer

By using this batch-like process, it is possible to save precious time while the robot does not have to do the whole test cycle for every single cover. This method works best if there are enough covers before the sprue breaker available to fill the buffer, else the robot must wait for new covers to arrive in the spruebreaker module.

For this application the sensor E2K-C from manufacturer Omron is chosen, which is a capacitive proximity sensor. The sensor is placed after the spruebreaking module. Every cover which rolls down is sensed and the buffer counter is raised with one. If the counter reaches 6 then the robot stops breaking sprues and start the quality check. The normally closed option is here also used to detect is the sensor is broken. If the cover approaches the sensor, the capacitive proximity sensor switch opens and the output signal of the sensor becomes low. A counter for the amount of covers in the buffer can be increased with one. The cover rolls further and the sensor signal becomes high again (+24 V). The data sheets and the decision list can be found in appendix VI.

4.6 WEIGHING DEVICE

Since the covers already roll in a rail coming from the sprue breaker, they might as well be rolled onto the weighing scale at once. Doing the weighing after the quality check results in an extra robot movement. One movement can be saved, if the covers rolls by itself on the weighing device. Therefore is chosen to do the weighing before the quality check. All cover data must be stored, but the data of the rejected covers must be marked when it is saved in the production log file.

The covers rolls down into a pre shaped rail, which is mounted on the weighing device. When the cover rolls on the weighing device, the output of this weighing device changes. This can be used as a trigger for the computer program to store the weighing data. The program must also send a signal to the robot. The robot can then pick the cover from the weighing device to the quality check. This weighing device can be an electronic balance, or just a sensor, see figure 24a. The choice of the two solutions depends on the way the assembly line is controlled. If a personal computer is used, the weight can be converted by a sensing element to a voltage value. This voltage value can be interpreted by the computer and displayed on the screen. In case of using a transducer, inductive or reluctance load cells can be used. Their capacity range varies from 5 grams up to thousands of kilos, furnish moderate to high accuracy's and their repeatability accuracy is approximately 0.05%.

A solution for weighing the covers is to use a load cell which gives a voltage reading as output i.e. 0-10V corresponds with 0-100 grams. The hardware board reads the output of the sensor and software converts this signal to an actual weight. Some sensors are not linear but this can be compensated with software.
An electronic balance is far more expensive, but gives easy to process data. An accurate balance is most of the time built in a box, and it can be more difficult to integrated an electronic balance in the process line.

Weighing will be done with a load cell and not with an electronic balance. A load cell is a transducer that converts a load acting on it into an analog electrical signal. This conversion is achieved by the physical deformation of strain gages which are bonded to the load cell beam and wired in a Wheatstone bridge configuration. Weight applied to the load cell either through compression or tension produces a deflection of the beam which introduces strain on the gages. The strain will produce an electrical resistance change proportional to the load. The sketches of the weighing device can be found in appendix V.

Leveling operations are important for the accuracy of the system. These include horizontal leveling and vertical positioning of the weight. It is important not to overload the cell during the mounting procedure. Even very transient overloads can damage a cell. To prevent this, some dummies during the installation can be used.

The suitable load cell for this application is found in an Omega catalog. A 50 gram load cell is enough to weigh the cover but there must be built something to let the cover rest. This weighs probably more than 23 grams so a 50 gram load cell in not suitable. The next weigh limit after 50 grams is 500 grams. The output of the load cell is attached to LabVIEW. The found specifications of the load cell can be found in appendix VI.

4.7 QUALITY CHECK

The test site can be divided into two sections, the roundness check and the clarity check.

4.7.1 The roundness check

The robot picks the cover up at the top, from the balance and holds it against a wheel and turns the cover around, see figure 24b.The wheel is attached to a little bar and the bar is attached to a potentiometer. If there are any irregularities, the potentiometer attached to the little bar senses this and the cover must be rejected. After the test, the cover is, depending of the results of the tests, moved to the rejection box or to the packaging table. The robot returns to the balance if there is a cover. If there is not a cover available, the robot moves to the sprue breaker and start breaking of the sprues from the covers in the spruebreaker module.

The cover rolls in the rail, placed under an angle, towards the test site. A little blocker in the rail stops the rolling cover at a certain point. The cover rests on two rollers and one of the rollers is attached to a motor. The motor starts running and causes the cover to turn. A photosensor is mounted to the module and emits a light to through the cover upon the receiversensor.On the edge of the cover the light is blocked and the light does not reach the receiever. If there are any irregularities, like cover is too small or not round, more light shines on the recieversensor. If this happens the cover must be rejected.

Figure 24a Weighing Figure 24b Roundness check

Another setup checking the roundness is rolling down the cover in a rail. A round cover has its specific characteristics. An acoustic sensor can listen to the rolling covers and decide whether the cover is round or not. If a cover is rejected, the cover was not round.

The first solution is the most suitable solution for this project. The potentiometers are relativity cheap and uses a different method to check the roundness. The robot is easy to program and is accurate compared with a home made device.
For the second solution the robot is not needed, because the cover rolls in the test site. This solution contains a lot of moving parts, like a motor, actuators and rollers. This means that the possibility something can go wrong is increased compared to a solution which contains less moving parts or power needing parts.
In the third solution the covers have to roll down which means loss of height. Acoustic sensors are sensitive and they can register also the surrounding noises and vibrations. These noises and vibrations can lead to rejection of good covers.

The little wheel is hold against the edge of the plastic cover and starts turning. It scans the surface of the edge and if the cover is not round, the wheel follows and moves down. The down movement is noticed by the potentiometer due to the little bar which is attached to both. A little bar is better because then even little changes on the surface of the cover can be detected. The current through the potentiometer is constant so when the potentiometer twist a little, the voltage over the potentiometer change, see figure 25. The voltage is measured and sent to LabVIEW. The voltage of the potentiometer is displayed in a graph. If the voltage exceeds a certain value, determent by the operator, the cover must be rejected. The weigh data and the quality check data must be marked as rejected and saved in the production log file.
Figure 25 Electrical schema and method of roundness check
4.7.2 Clarity check

If the plastic pellets contain moisture, there appear airbubbles in the cover. These covers must be detected and rejected.

The cover is held by the robot hand and positioned for the sensors. The robot hand starts to spin in front of the sensors. The sensor receives light from a light source and when an airbubble appears between the light and the sensor, the light is reflected. Less light arrives in the light sensitive sensor so the sensor gives a lower signal value that can be interpreted as a non clear cover.

The cover is placed on the lens of a CCD camera. A light shines upon the cover and the CCD camera transforms the signal to a gray tone. The receiver is built out of many cells. Every single cell is able to store a gray tone depending on the amount of light they were exposed to. The 'picture' of the new cover must be compared with a stored 'picture' of a cover without errors. An algorithm is used to detect the errors, and depending on the result of the calculation the cover can be rejected. The decision making is suitable LabVIEW task and it is a good and accurate method.

The first solution is here the final solution for this project. Since both clarity and roundness are checked spinning the cover, so they can be checked at the same time. This means the robot has to spin the cover only once. Simultaneous data readings means that incoming data must be sampled, but that is not a problem if you spin the cover not to fast. The second solution is a more accurate solution but again the CCD camera is not available in the laboratory and it is quite expensive.

For the clarity check another sensors are necessary. The cheapest sensing method for the clarity check is with photodiodes. The lights is transmitted from one side through the transparent cover and the light reaches the photodiodes receivers see figure 26. In the dark areas it is not possible to check for airbubbles. The light source are infra red LED's, and they are placed before every single photodiode. The sensors placed on the right side of the cover, scan the areas where the left sensors can not scan, so the whole cover is checked.
Figure 26 Method of scanning and scanable areas

Another solution for the light source is a normal light bulb. But there are some disadvantages like: the lightbulb ages and transmits than less light, the lightbulb consumes more energy, the whole setup must be placed in a dark place, were no light can enter. Because the receiver is sensitive for the whole spectrum of daylight.

A infra red light receiver has its peak sensitivity around 800 nm, in the infra red spectra. For this solution there is no need to cover the whole setup, while the test can be done in normal day light. If the light amount differs then the cover contains airbubbles. Covers where this occur, must be rejected from the process line. The weigh data en the process data must be marked and stored in the production log file.

Photodiodes are available in many different forms and sizes. From very small receivers 0.75 mm² to larger ones 100 mm². The receivers also differ in shape, round is the most common but rectangular receivers are also available. For our application is no need for big receivers, because the resolution of these receivers is not good enough for this application. When the receiver is small, the receiver can better detect the smaller airbubbles.

Figure 27 Electrical schema transmitter and receiver for clarity check

The transmitter part consists out of 10 parallel linked infra red LED's with a diameter of 5mm. The peak wavelength of the LED is 940 nm, further information about the LED can be found in appendix VI. The current through the LED's is reduced by the resistant R=2.2k to 11mA.
The second part of the schema is the receiver part. With the potentio meter P1, the voltage on the + of the opamp can be adjusted to a certain level. If the photodiode receives more IR light, the output voltage of the opamp lowers and if the photodiode receives less IR light the output voltage of the opamp raises. With the resistor R1 the current is kept low. The diode D2 is used to filter the voltage which is lower than the adjusted voltage. If the receiver detects an airbubble the voltage raises and after the resistor R1 a positive voltage is present. The signal is measurable at the end of the schema and is connected to LabVIEW and LabVIEW must reject the cover. All the other receivers are connected to the two OR chips. The receiver diodes are from Siemens and coded as BPX65. The diameter of the light sensitive area is 5.3 mm and the peak of the photosensitivity is 850nm. See appendix VI for the datasheets.

4.8 PACKAGING THE COVERS

After quality check, the approved covers are placed by the robot in square groups of four on the packing table. The covers have to be packed manually in packages of four. In the near future, this part of the assembly line can also be fully automated, but that is not part of the project yet.

4.9 THE GMF S110R ROBOT

In some of the solutions for the assembly line modules described in the previous paragraphs is a robot mentioned. One robot, the GMF S110R, is available and usable for the tasks that are mentioned in these paragraphs.

The S110R is an articulated robot with six axes, that is mounted to the floor in the middle of the laboratory. The robot has its own controller that acts like a PLC; it is programmed with a ladder logic language that consists out of three different types of instructions. G-codes are used for moving the robot, F-codes control the motion speed of the robot and S-codes are used for I/O signal control of the controller. The S-codes can be used to let the robot controller communicate with other devices, like PLCs, and to enter different branches of program code depending on the input channels. The logic levels for the I/O channels of the robot controller are 0/+24V.

The maximum load capacity at the wrist is 10 kg. This is for our process not really necessary, but the other smaller robots are either slower or their reach is too small.

One of the most important parameters of a robot is its position repeatability, which reflects the accuracy of the robot positioning. The repeatability of the S110R is 0.2mm. More specific technical data about the robot can be found in appendix VII.

A main problems to overcome is the fact that the robot is located about 10 meters from the Injection Molding Machine, so some kind of transportation has to be designed to move the covers from the IMM to the robot. This is discussed in paragraph 4.10.

4.10 OVERCOMING THE DISTANCE BETWEEN ROBOT AND IMM

One of the main problems of the process flow is the big distance between the IMM and the S-110R robot, about 10 meters.
Figure 28 Location robot and IMM in the laboratory.

Neither the robot or the machine can be moved. This distance results in a big time constant.

Several solutions for this problem have been found:

Transporting the covers with an automated guided vehicle (AGV) from the IMM to the processing area.
Using a 10 meter conveyor belt.
First transporting them vertically, then sliding the covers down from a ramp. Tests have been done with different angles and different surfaces. The smoothest surfaces proved to be Formica and stainless steel.
Since the covers are round, rolling down the covers is a possibility.

The AGV is available in the laboratory, but it is fairly slow. The AGV can not make it to and from the IMM in the time one cover is produced, and this results in batch delivery. Doing batch deliveries, fast feedback to the operator about the production is impossible, and some process modules can get serious difficulties if they have to process several covers at once.
The 10 meter conveyor belt is simple to control and the most simple solution. The disadvantage of this solution is the costs of such a conveyor belt.

The sliding solution is not expensive but Formica still needs an angle of 15° to let the covers slide out of a static position. A steel surface needs about the same angle, but to make the entire ramp out of stainless steel is pretty expensive. The 15° angle causes a far too high start of the ramp around 2.75 meters above the ground, so this solution is not very practical.

Rolling covers down seems to be the best solution to transport the covers for this project. The covers have not too much friction, so they roll with only a few degrees decline. When the covers are placed between two C-shaped rails, they can easily roll down with approximately a five degree decline.

4.11 THE TOTAL PROCESS LINE

Having discussed all the different process stages, a total overview can now be given (see figure 29). The covers fall out of the IMM into the gutter that aligns them. Then, the bad covers get taken out of the flow in the filter. The elevator is needed to gain some height, allowing the transport rail to roll the covers down. Thereafter, the covers enter the sprue braking station where the sprue gets broken off. The 6-piece buffer is necessary to allow the robot to work in series instead of "one by one". At the quality check station, the covers are being weighed, and checked on clarity and roundness. The approved covers are then placed on the packaging table in groups of four.

Figure 29 Total process overview

4.12 PROCESS CONTROL

Most of the process electronics are just switches, leading to boolean decisions. Therefore, the main part of the process can be controlled by ladder logic, which means that a Programmable Logic Controller (PLC) is able to control the process, except for the weighing of the covers. A Siemens simatic CPU- 214 PLC is available for the project. Data storage from the process and easy accessibility of data and displaying of this data is not the best point of a PLC.

The requirements for remote process monitoring and data logging lead to the use of a Supervisory Control And Data Acquisition (SCADA) application. Basically, this is a computer program that can monitor and control the whole process using a data-acquisition hardware board, and store the acquired data in computer files. A SCADA application can do everything what a Programmable Logic Controller (PLC) can do, and much more.

The SCADA application must also be able to do calculations on collected data, like quality check data. In the laboratory are LabVIEW hardware boards and LabVIEW 4.0 software available. LabVIEW is capable of controlling the whole process, doing the data-acquisition, and display the collected data on a computer monitor in a user friendly way. However, for the educational part of the process, the PLC is used for the boolean control of the process, leaving LabVIEW to do supervisory control collecting, processing, displaying and storing the data. LabVIEW makes the decision if a cover passes the quality check or not. This means that the PLC and LabVIEW need to communicate.

The S-110R robot communicates with the PLC and not with LabVIEW, because they have compatible logic levels (0/+24V). Connecting the robot to the computer running LabVIEW requires a special interface, and the PLC can do the communication just as well as LabVIEW.
The LabVIEW hardware boards has only a few digital TTL I/O channels. TTL is 5V logic and to connect it with the robot (24V). The cheapest way to solve this, is to put a driver between the two devices like figure 30. Another, much more expensive, way to solve this problem is to use a component from the National instruments catalog. The device can convert from TTL inputs to 24V outputs.

Figure 30 Driver for conversion from TTL to 24V.

Figure 31 Device interfaces

The different device interfaces for data transfer and data interchange in the process are given in figure 31.

This is an internal computer communication channel. Alarm signals that are acquired by LabVIEW have to be saved in a computer file.
This communication is done by NI-DAQ software which allows LabVIEW software to communicate with its hardware.
An analog signal from the weighing devices sensor is sent to the LabVIEW DAQ board, measuring weight.
LabVIEW measures three temperature zones and a pressure in the IMM (analog signals). The sensor at the outlet of the IMM indicates if a cover is produced (switch signal). LabVIEW has also the ability to start and to stop the IMM .
This connection is used for the communication between the PLC and LabVIEW. The protocol for the communication is described in paragraph 4.11.5.
LabVIEW must know when to start the data acquisition for the quality check. When the robot is in position, it gives a signal by making an logical output channel high. This signal must be sensed by LabVIEW on an input channel and triggers the data-acquisition.
Two switch-signals from the IMM (Mold open/closed B17 and Outlet switch B10) are sent to the PLC so that the PLC can time if a part actually fell out of the outlet after the mold opened. Both are digital signals.
Some of the sensor signals are sent to LabVIEW, allowing LabVIEW to calculate how many covers were rejected, produced, packaged, etc. These are boolean signals.
The PLC controls the assembly line, so it has to be provided with all the switch-signals from the process. All these signals are boolean.
The PLC communicates with the robot; it tells the robot whether or not to do clarity checks, if the robot has to do sprue braking or quality checks, when to start, etc. These signals are all boolean. The PLC and the robot have the same logic levels, so no interface is necessary .
The PLC controls two actuators; One controls the trapdoor of the filter, the other the blocking pin at the end of the buffer. Two digital PLC output channels are needed.
The DC rail lift motor is activated by a relay that is controlled by a PLC digital output channel.

4.12.1 Sensors of the process line

This is a summary of the sensors purposes and the sensors have the same number as in figure 29:

The first sensor is a switch, which is already installed in the IMM, and officially coded B17. The switch senses if the mold is opened or closed; if the mold is closed, the switch is closed. This sensor also starts a timer. If this timer is not stopped within 3 seconds by triggering sensor 2 , then no cover was produced by the IMM. This can also mean that during the previous cycle a sprue got stuck in the IMM. This serious malfunction must be reported to the operator as fast as possible, since the IMM is then no longer capable of producing covers if the sprue is not manually removed.
Sensor two is also an already installed switch (B10) and attached to the outlet of the IMM. If a cover falls through the outlet into the gutter, the switch is closed for a fraction of a second. sensor 2 is used to stop the timer that was started by sensor 1.
This is a sensor which is used to check if the covers lay to long in front of the filter. The signal is sent to the process controller and every time a cover passes the sensor beam, a timer has to be started. If the timer exceeds 3 seconds, the PLC activates an actuator that triggers the trap door under the filter.
This sensor must sense if a cover falls out of the process. The sensor is mounted under the turner and when a cover is too small, it falls out of the rail, through the sensor beam. The timer which is started by sensor 2 must be stopped when this sensor is triggered. The counter for cover too small must be increased with one.
This sensor is placed to keep track of the covers in the process line. The covers are able to travel the first process section between sensor two and sensor 5 within 10 seconds, so the timer that was started at by switching sensor 2, must be stopped when sensor 5 is triggered. If the timer exceeds 10 seconds, the operator must be warned. At every rejection of the cover at the filter module the timer must also be stopped. Sensor 5 also starts a new timer for sector 2.
The signal of this sensor stops the timer that was started by sensor 5. The purpose of this sensor is to sense if a cover got stuck in the transport rail. If the timer exceeds 15 seconds the operator must be alarmed.
The sensor in the sprue breaking module is necessary to detect if there is a cover in the sprue breaking module. The robot can now be notified that there is a cover available.
This sensor is used to count the number of covers in the buffer. If the counter has count six covers, the robot has to change its task to quality checking. The robot moves to the balance and starts the quality check.
The balance sensor measures the weight of the covers. The weighing data are stored in a special weighing file. After the cover is weighed, within 1 second, the robot picks up the cover and moves to the quality check module.
The edge of the cover is scanned by a wheel. The sensor data is checked on its quality limits determent by the operator. If the signal of the sensor is too small, the cover must be rejected.
During the same test, sensors are used to check the clarity. If the signal received by the sensor is not right the cover must be rejected.

All the sensor specifications can be found in appendix VI.

Besides these sensors, a few sensors are needed to measure three different temperatures on the IMM and injection pressure. Two heat zones are already measured in the current setup with two thermocouples. The other sensors need to be installed.

The letters A, B, and C indicate assembly line parts. The type of actuators and elevator is not determent yet, but later when the modules are built.

This is the actuator that opens and closes the trap door under the filter.
The elevator is driven by a DC motor with constant speed.
Actuator controls the blocking pin of the 6-piece buffer.

4.12.2 The Programmable Logic Controller (PLC)

In the laboratory is already a PLC from Siemens available, type SIMATIC CPU 214. The base unit has 14 discrete inputs and 10 discrete outputs. Two type of expansion units are also available; One has another 8 discrete outputs, the other has one 12 bit analog input and three 12 bit analog outputs. If even more I/O channels are needed, more expansion units (up to a maximum of seven) can be connected. Power supply and logic levels are 0/+24V. The program is not lost when the PLC is power cycled, because the memory device retains its contents without the application of any power. The data is maintained by a super capacitor which last 190 hours (typical) and there are no batteries required. The fast boolean execution takes 0.8 microseconds per instruction and has also a real-time clock.

4.12.3 LabVIEW

LabVIEW, or Laboratory Virtual Instrument Engineering Workbench, is a graphical programming language that has been widely adopted throughout industry, academia, and research laboratories as the standard for data acquisition and instrument control software. Running on Macintoshes, Sun SPARC stations, HP 9000/700 Series workstations, and on PC's running Windows 3.1, Windows NT or Windows 95. LabVIEW is a powerful and flexible instrumentation and analysis software system. Computers are much more flexible than standard instruments, and creating your own LabVIEW program, or Virtual Instrument (VI) is simple.

The LabVIEW program is going to be a main part of this project, so it must be treated as a project within a project, and developed with the same System Development Method as the entire process line. This can be found in chapter 5.

4.12.4 Emergency stop wiring

At all the different process line modules must be an emergency stop button present that shuts down the whole assembly line immediately. The LabVIEW front panel must have such a button.

Figure 32 Bad designed emergency stop system

Figure 32 shows a bad design of an emergency stop system. For example, In case of a PLC malfunction, LabVIEW is not able to shut down the robot. To prevent these kind of problems, the wiring must be done as shown in figure 33.

No voltage available means all devices have to stop their action. This setup prevents that a wire fracture can let the system malfunction in case of an emergency. 24 volts is used in consideration to compatibility: the robot, the PLC and the sensors all use 24V power supply.

Elevator motor power and IMM power are directly cut in case of an Emergency stop. The robot controller gets a low signal on its ESP terminal, which results in an immediate stop. The PLC outputs must be made powerless and go into a save mode for the process. The inputs of the PLC keep their power. When the robot has to be restarted, the Reset terminal has to get a signal from the PLC. The LabVIEW hardware gives a signal on a digital TTL channel, and the PLC receives its signal on a digital input (DI).

In case of an emergency stop Al signals are isolated by either thermocouples or relays; the elevator motor power and the IMM are connected via relays because they have to switch power.

Figure 33 Emergency stop wiring setup

4.12.5 The PLC-LabVIEW communications channel

LabVIEW has to communicate with the Siemens PLC which controls the actual assembly line. Information has to be sent in both ways; errors detected by the PLC must be reported to LabVIEW. The PLC must know when to start the process, if the robot has to do clarity checks or not, and if a cover is approved or rejected during the quality check.

Communication can be done with a number of digital I/O lines. Each message has then its own unique bit pattern. However, this requires several digital I/O lines, and the digital voltage levels of the PLC and LabVIEW are not compatible, and an interface is needed.

The PLC has also the ability to communicate over a RS-232 serial communication port.
RS-232 (ANSI/EIA-232 Standard) is used for many purposes, such as connecting a mouse, printer or modem as well as industrial instrumentation. Due to improvements in line drivers and cables, applications often increase the performance of RS-232 beyond the distance and speed listed in the standard. RS-232 is limited to point-to-point connections between serial ports and devices.

The standard only specifies the hardware protocol, the electrical characteristics of the communication link. Parameters as number of data-bits, parity and baud rate are specified by the PLC and LabVIEW software.

RS-232

type of transmission lines Unbalanced

Maximum cable length in meters 15

Maximum data rate 20 kb/s

Maximum number of receivers/drivers 1/1

Table 3 RS-232 standard specifications

The RS-232 interface standard specifies only full duplex communications. This means that the interface consists out of separate lines for transmitting and receiving. With separate lines, transmit and receive operations which can take place simultaneously, without danger for data-collisions.

RS-232 is specifically designed for communication using asynchronous serial protocols. These protocols are typically ASCII based. Communication with devices is done by sending strings or characters to and from the COM port. There is no danger that the receiver misses data if it is busy at the time with receiving something else, because the data is written to a buffer.

Serial communication is also possible with LabVIEW, so all communication with the PLC is done with this serial port by sending single ASCII-characters. This enables sending numerous different message-codes, saving several digital I/O lines compared with the digital TTL lines solution. Another disadvantage of the TTL is sending data over long distances (a few meters). The TTL line become very sensitive for interference. When drivers are used the TTL lines are less sensitive for interference.

More information about the PLC and LabVIEW can be found in appendices VIII and IX.

The next settings are chosen for serial I/O:

Baud rate 9600. The PLC is able to send and receive data at a rate

between 2400 and 19200 Baud. LabVIEW is able to do the same, but communication tests between the PLC and LabVIEW have proven that the error rate getting larger, so the Baud rate is set on 9600.

Parity Even. The parity is used to check if the received data is

correct. The second way to see if the send data correctly is received is to answer with an acknowledge.

Number of stop bits 1.
Number of data bits 7. The ASCII characters which are used have a lower value

than 128 (2⁷). To send 8 bits is not necessary at all.

Handshake mode None

5 LabVIEW

5.1 DEFINITION STUDY

5.1.1 Project

An industrial assembly line containing logic switches, actuators, a balance and sensors has to be controlled by a combination of a Programmable Logic Controller (PLC) and LabVIEW Virtual Instrumentation software. The PLC controls the boolean part of the process, while LabVIEW is the user interface, collects data and executes calculations on it. This sub-project contains the design and development of the LabVIEW program.

The assembly line that needs to be controlled is described in the previous chapters of this document.

5.1.2 Planning

The LabVIEW programming must be ready by May 23, 1997. After that date, the project is taken over by a new team.

5.1.3 What is LabVIEW?

LabVIEW is a programming environment in which programs can be created with graphics. In this regard it differs from traditional programming language