The intent of this project is to design and construct a machine to automate the placement of strand in the EMABOND^TM process for Ashland Chemical.  This process utilizes a metal impregnated thermal plastic, EMA^TM, to bond tongue-and-groove plastic parts via induction welding. Several designs were considered in order to meet all of the design specifications.  The team members and advisor evaluated these design concepts.  For the final design an x-y positioning table will be used to control a fixtured part while a single head device will be used to place the EMA^TM strand.  This head will be suspended above the part on a gantry support.  The controls for the head will incorporate three-directional motion and 360 degree rotation.  Additionally, the head will contain all systems (heater, cutter, and sensor) necessary for accurately placing the EMA^TM.

After Ashland’s approval of the final design concept, drawings were drafted and various components were researched.  Following the design phase of the project, the team tested the various components of the head design while constructing and troubleshooting the device.  Due to budget constraints, Ashland decided to only have the group build a manual prototype device.  This device will be automated using the specified drive control system at a later time.  The machine will be constructed to reduce labor costs of the current process, incur zero rejects, and interface with existing production methods and equipment.  These objectives will be accomplished with a machine capable of high production rates utilizing a spatially efficient and safety conscious configuration.  The tasks to be completed for this project include a proposal, progress presentation, project website, educational poster, and final report.  A cost analysis determined that the project for the second semester costs the following: $17,500 for labor, $17,500 for labor overhead, $1,000 for machining costs, $1,535 for materials, and $230 for material overhead for a total of $37,765.

An animation of the final design demonstrates how this machine will operate once automated.

Background

This project is being done under the auspices of Ashland Chemical.  Ashland has specific interests in petroleum based industries and products; hence it manufactures a multitude of products that are chemical derivatives of petroleum. The multiple industries of Ashland include the Ashland Specialty Chemical division.

One division of Ashland Specialty Chemical is the Specialty Adhesives & Polymers (SPA) Division.  The EMABOND^TM System, which our project will enhance, is one of SPA’s innovative technologies.  This technology involves the joining of thermoplastics, which is limited to a few techniques each with their own drawbacks and limitations.  The EMABOND^TM System utilizes the technique of joining thermoplastics at a bond line.  Welding at a bond line requires the plastic to reach melt temperatures, which is accomplished in many ways.  One process is sonic or vibration welding which relies on mechanical friction between two parts to create the requisite heat.  An alternative process is hot plate welding which uses a heating element to soften a bond area and mesh it with another softened part.  The EMABOND^TM process uses induction welding to heat a weld material within a joint.

There are numerous advantages to the EMABOND^TM process.  One advantage is the ability to weld parts quickly.  The rate of welding usually fits within the time frames of high quantity production lines, which gives industries another reason to use the process.  It also enables the welding of parts containing reinforced fibers and produces joints that still maintain the property of the material.  The use of glass-reinforced plastics is growing in industries such as the automotive industry because the use of a plastic with high strength and low weight is necessary to reduce costs. Previously, molding large parts was difficult due to the mating of the components.  The EMABOND^TM process is not hindered by the increased size of a part.  Although welding along different planes is not favorable, this process can be adapted to complicated pieces.  These aforementioned applications previously relied on mechanical fasteners, whereas the EMABOND^TM process produces welds with strengths equivalent to a single structure without a joint, which enables the creation of pressure-tight welds and hermetic seals.

Based on inducing heat within the joint, induction welding uses the magnetic properties of ferrous metals in conjunction with a thermoplastic compound to create a weld.  Each part specific machine that performs the welding process is composed of a few basic components.  These include the induction generator, a copper work coil, a fixture, and the EMA^TM material of proprietary composition that is placed in the joint of a product.  The induction generator produces energy in frequencies capable of exciting the ferromagnetic particles within the EMA^TM material through hysteresis and eddy currents.  Shaped in the outline of the product, the copper work coil surrounds the part at the bond line and transfers the electromagnetic energy.  The fixture surrounds the part and ensures stability and adequate support of the bond area.

The typical process, illustrated in Figure 2, requires the EMA^TM material to be placed in the groove of a tongue and groove joint.  The two mating parts are then joined in a fixture with a pressure plate to ensure pressure between the two halves.  A work coil is affixed to a nest in the setup.  The coils are then energized and as the EMA^TM material becomes molten the parts are pressed together.  During welding the surrounding material of the joint begins to melt.  The EMA^TM material then fills the joint as the pressure reaches a maximum and the part is fully fused.  There are various joints that the EMA^TM process can weld.  The preferred one is a shear joint composed of a tongue and groove configuration.  A strand of EMA^TM is placed in the bottom of the groove and the tongue is placed on top of the EMA^TM.  As the parts are joined, the EMA^TM flows to fill the space of the joint.  This is the optimum design due to the strength of the joint, which is principally a factor of the large surface area of the shape.

The most important component is the proprietary EMABOND^TM material.  The EMA^TM material can be made in various forms for different applications.  It can be created in strand, tape or injection molded shapes of varying lengths.  The formula generally consists of a set of thermoplastics that resemble the composition of the parts to be joined.  The best results are achieved between plastic parts with similar melt indexes.  The particles within the thermoplastic can be either iron or stainless steel of various sizes and concentrations.  These properties are adjusted based on the application.  Different compositions of material as well as various shapes and sizes are manufactured to suit the product application.  This customization maximizes the speed and efficiency of the process.
 

Current Process

The time needed to place the EMA^TM depends on the length and shape of the bond line.  A longer and more circuitous bond line will require greater time to place the EMA^TM strand.  Another factor that can hinder strand placement is the size of the strand.  If the strand is thin it will be more difficult to handle.  The groove can be undersized or oversized depending on tolerances of the plastic molded parts.  If a strand is much smaller in diameter than the groove it can become displaced during placement.  Conversely, if the strand is too large for the groove, a tight fit can hinder strand insertion.

Manual insertion of the strand is not a challenging task, however placement can be slow when compared to automated production lines.  The need to make a repeatable, accurate and fast way of delivering EMA^TM to the groove is necessary for companies to quickly produce parts.  If the speed is enhanced via machine then production rates will increase and costs will be reduced.

The manual placement of the strand begins with the part on a table (Figures 4 & 5).  The laborer needs a strand of correctly sized EMA^TM and a tool such as a screwdriver or a pick (Figure 6).  The EMA^TM is held in one hand over the groove while the other hand holds the tool, which pushes the end of strand into the groove (Figure 7).  The pick is then used to guide strand into the groove (Figure 8).  As the pick follows around the groove the strand is forced into place (Figure 9).  Both hands are needed to place the strand, which can result in difficulties due to the part merely resting on a table.  As the laborer puts pressure on the sides of the part it can move, thus slowing the placement of strand.  A part secured in place would ensure no movement or rotation.

 

 

 

 

 

Previously Explored Alternatives

As previously stated, the process of placing strand into mating grooves has primarily been done by hand.  Certain manufacturers have attempted to automate their processes.  One manufacturer that dealt only with circular parts utilized a rotating table to spin the part.  However, an assembly worker was still required to place and guide the strand into the groove as the piece rotated.  Johnson and Johnson, Inc. explored the placement of similar bond material in the production of packaging products.  This effort yielded an extremely complex and large machine that was intended to completely automate their process, but the configuration was plagued with mechanical difficulties and loss of production.  It was eventually abandoned and more assembly workers were implemented in order to achieve the higher volumes of production that were desired.

Automation of assembly processes is utilized by many industries.  One aspect of automation incorporates robots or robotic devices.  Applications for robots include welding in the automotive industry, handling high temperature components in foundries, and dipping and coating pieces in quenching and plating operations. Although the EMABOND^TM process does not exist in such extreme environments as the welding and chemical plating operations, some of this existing technology may be used in developing the final design.  Multiple heads are utilized on a single robotic platform to increase production rates (Figure 11).  Gripping methods, guiding, and control are all robotic areas that can be applied to the EMABOND^TM automation device (Figure 12).

 

Method of Approach: Construction and Testing

The main requirements to achieve the project goal are to accurately place and seat the strand in the groove, while ensuring the start and end points are aligned with no overlap to within one half diameter of strand.  Last semester focused on the design phase of the project.  To accomplish the construction of the device this semester, test plans for the major components have been developed.  Once the heater and sensor are tested, the prototype will be assembled.  This prototype will be tested and debugged prior to integration with the motor control system.  Further testing will be completed once the system is motorized.

NOTE:  The Appendix  contains all information pertaining to the fall semester’s schedule.

The project is divided into several stages and scheduled accordingly throughout the semester.  The stages consist of obtaining and testing the individual components, machining the gantry and head designs, assembly, and testing of the entire device.  Once the prototype device is tested, adjusted, and finalized, the x-y table can be purchased and integrated with the system.

The project description that follows refers to the tasks on the Spring Semester Gantt Chart in Appendix A.  The numbers in parentheses correspond to the ordered tasks on the Gantt chart.

The spring semester focuses on the construction and testing phase of the project.  Although specifications for all components were created last semester, this semester began with finalizing these specifications (2).  After meeting with Ashland Chemical and our project advisor following last semester’s final presentation, there was a need for drawing revisions (3).  Once all specifications and drawings are finalized, parts and materials could be ordered (4).  Most of the semester will be spent constructing the prototype.  This consists of having a machinist build the customized parts and test fixture (10, 11).  Once all parts are machined and ordered components are received, the gearbox and gantry can be assembled (12).  Each purchased component will need to be tested individually according to the test plans developed last semester.  The heater (13), sensor (14), computer program (15), and drive mechanism (16) will all be tested individually prior to being integrated into the prototype assembly. Following the component testing, the prototype device will be assembled and tested (17) to determine if the device meets the design specifications set forth by Ashland.

During the course of the semester, Ashland Chemical decided to limit the project.  Due to budgetary constraints, the device will not be integrated with a drive and control system.  The project objective has been modified to require construction of a prototype head device and its gantry support.  This device will be tested manually to prove its effectiveness.  Once the manual device has been proven to work, it will be automated using the researched drive and control system.  As a result, all tasks involving the x-y table and drive control system have been eliminated from the project’s tasks.
 

Heater Test Plan

The basic plan for testing the heater will be to simulate the strand placer’s head by manually feeding the Ema^TM through an aluminum test fixture with the Watlow Band Heater running at the 97.4 Watts determined from the calculations.  The aluminum test fixture will be the same size as the actual prototype head, with a .5” outer diameter and a .0625” inner diameter.  The Ema^TM will be manually fed through the test fixture at approximately 6 inches per second in order to examine the effects of the heat on the Ema^TM at the outlet.  The Ema^TM must be pliable at this outlet temperature, but neither too soft nor too stiff.  If the desired physical properties of the Ema^TM are not obtained, the wattage of the heater will be adjusted accordingly until the Ema^TM reaches the desired pliability for effectively rounding sharp corners in a part.

Ema^TM was originally tested using an oven.  It was determined that 120 degrees Fahrenheit was the necessary temperature that was needed for the Ema^TM to possess the pliable properties that would enable it to successfully fit around sharp corners in a plastic part.

An analysis was performed on a model of the heater using Ansys.  Based on the model that was produced, the heater appeared to have the necessary power in order to heat the Ema^TM to the required temperature of 120 degrees F.

Next, heat transfer calculations were performed in order to determine the amount of time that would be necessary to heat the Ema^TM to the required temperature of 120 degrees F.  All property values were found using Heat Transfer: A Practical Approach by Yunus A Cengel.  The Ema^TM properties were estimated as half stainless steel and half polypropylene due to the fact that Ashland was unable to provide the actual properties of the material.  As can be seen from the calculations in Appendix F, the required time would be about 237 seconds, or roughly four minutes.  This is obviously insufficient.  Therefore, in order to greatly reduce this time, it would be necessary to add more heaters on to the head and/or provide some sort of warm air convection within the head in order to heat the Ema^TM more quickly and make it more pliable.

Unfortunately, the heater and thermocouples are no longer working, and it will take more trouble-shooting in order to fix them and to test the Ema^TM properties using the heater on the head.  This will verify the calculations and modeling results and allow us to better understand what will be necessary in order to heat the Ema^TM to 120 degrees F more quickly.

Sensor Test Parameters

The sensor must be tested for its accuracy in detecting the strand within a groove.  It must also be tested to determine what conditions, if any, result in a false reading.  To simulate the conditions of actual operation the sensor is mounted on a milling machine with a mock groove placed on the moving table beneath, as illustrated in Figure 13.  The sensor is fixtured in an aluminum support piece.  When the table is moved under the working sensor it is capable of simulating the motion of the x-y table.  Once the sensor beam is aligned directly above the groove, the table moves and the beam travels along the path of the groove.  When the sensor beam is above the groove it does not “sense” the strand.  The sensor is attuned to the height of the beam so that the sensor does not “see” anything until the strand is present.  When the strand is present the sensor will turn on a red indicator light.  This test replicates the action of the fully automated system.  When the sensor is activated it will not see anything until the head device returns to beginning and the first piece of strand is detected.

 

Several trials of the test will determine whether or not the sensor purchased will meet the requirements of the strand placer.  All tests were done on the straight portions of the groove.  Ten trials were performed for each of three groove distances.  Each set of tests was setup to determine the accuracy of the sensor and it frequency of accurate signal output.  The trials represent a different length traveled over the groove.  The strand end is placed at the midpoint of the travel distance.  Distances of one inch, two inches, and three inches were tested.  The first detection distance is an indicator of a false reading or the sensor detection of something other than the strand.  Since the portion of the groove traveled by the sensor is straight there should be zero false detection.

As is shown in Table 1 the optical sensor in the prototype has a percentage of false readings.  These false readings are based partially on the distance traveled by the sensor.  Since the width of the test groove is 0.08” wide and the sensor beam is 0.05” wide there is a small margin for error.  The 0.015” space on either side of the detection beam is very small.  Inconsistencies in the groove can cause the false detection output.  The path traveled by the head must be accurately aligned so false readings do not occur.  Therefore, on a moving prototype this sensor could be troublesome based on the alignment of the head as it travels over the groove.  The slight amount of sway that can occur with moving parts of the device is partially responsible for misalignment.  Imperfections of the molded plastic part can also contribute to the false output.  Since it is not always straight as it lays in the groove, the strand can also be a problem.  If the strand is closer to one side of the groove it might provide problems for the sensor to detect.  There are no problems based on the similar colors of the strand and the groove.  Incorrect alignment is the largest source of error leading to detection of the groove wall.
 
 

A few false readouts per trial show proper alignment is absolutely critical. Once a false detection occurs the head can be aligned more correctly.  However, in a production device a more accurate sensor or an alternative method for detecting the strand is necessary to meet the design specifications.  Since the accuracy of the head movement is critical to the sensor operation its improvement would be most beneficial correcting erroneous sensor outputs.
 

Design of the Device

Design Requirements

The design requirements as stated by Ashland Chemical are:

• Must be capable of being automated
• Table top unit
• Multi axis capable
• Have the capacity to insert stand with circular cross sections from 0.080” to 0.125” diameter
• Strand will feed off of a standard reel
• Will handle polypropelene and polyethelene strand, and possibly nylon strand
• Capable of zero reject operation
• Safe to operate
• Lay strand up to 6 inches per second

Comparison of Existing and Proposed EMA^TM Insertion Methods

This machine will become an element in an automated production line for customers of Ashland Inc.  It will increase the efficiency and production rates of the EMABOND^TM process.  A comparison of the current method versus the proposed method is as follows:
 

Parameter	Hand Placement of Strand	Automated Strand Placement
Process Speed	Varies with operator	Up to 6” per second
Reproducibility	Operator error can cause rejects	Process capability of ~ 0 defects
Part Flexible	Operator learning curve between parts	Individual programs based on part numbers eliminate start up delays
Cost	Wage and benefits	Initial machine cost
Operator Error	Leads to gaps/overlaps at strand ends	Controllable gap distance of less than ½ strand diameter
Maintenance	None required	Routine maintenance/possible production interruption

Design Concerns

The primary obstacle of design is in its applicability to the customer and to the industry that manufactures the product.  Creation of an end product capable of solving the problem or filling a need must be the main goal in design.  It is also important to ensure that the design is conducive to the industry in which it will operate.  The particular trouble of design requirements can be related to several key areas including cost of the product, difficulty to manufacture the product, and maintenance of the product.

A concern of foremost importance is the manufacturability of a design. Complex geometries can be difficult to mass-produce, therefore the design of the strand placement device incorporates simple and basic shapes.  The choice of materials is a key factor in cost as well as in the mass production manufacturing process.  The strand placer design uses mostly aluminum.  This will keep costs low, as well as ensure ease of machineability.  Components requiring a greater resistance to abrasion are made of hardened steel to prevent frequent maintenance, thus decreasing future problems and long-term costs.  The choice of periphery components such as electronics can determine the overall cost of the machine and the shape of the design.  Any secondary components purchased from vendors must be able to meet the standards of the design in effectiveness and cost.

The goal of the design usually focuses around a problem to be solved.  An effective design must solve the problem with accurate and reliable actions and minimize the amount of failure.  Some concerns with meeting the design requirements are the speed of the device or its ease of use.  The characteristics that can make the product successful depend on the overall sense of quality and satisfaction that the task is completed.
 

EMAWELD Strand Placer: Concept Designs

The challenge of placing a strand into a groove has led to several design concepts. Design specifications provided by Ashland have essentially narrowed the designs to two variations.  The tabletop unit can either be based on x-y table control or a rotary control device.  Either a single head or dual head design can deliver the strand.  There is also a choice between moving the head and moving the part.  The two head designs are comparable in concept to the two drive mechanism concepts. The x-y table can be suspended over a part and move the head.  Alternatively, the x-y table can move the part while the head is suspended stationary above.  In a similar manner, the rotary table can either rotate a suspended head or rotate the part while the suspended head would place the strand.  In the situation where the head would be moving, the two head concept is more applicable.  This second head would remain fixed next to the part to maintain the strand’s position in the groove while the suspended head carries the strand around the part.  There are six possible designs, without accounting for the variations in the dual head concept design.  An evaluation of these concepts is based on two major systems.  The drive mechanism will either be x-y table or rotary table driven.  It will also be considered in conjunction with the option to move the head or move the part.  These criteria are considered as one major section, while the number of heads is considered separate from the drive mechanisms.  The splitting of the systems allows for careful unbiased evaluation.  Each design concept is discussed below and presented with figures.

X-Y Table
The applicability of an x-y table is obvious to this project.  Complex movements are necessary to meet the demands of the geometry of the plastic parts.  Therefore, a device that moves in a precise and predetermined path requires technology to be developed for the placement device.  Since x-y tables are commercial products, the reliability, safety and accuracy are reasonably assured.

In one design, the x-y table is suspended from a gantry and is used to carry a head above the molded part.  The part is secured in a fixture while the head mechanism operates above it.  This head would contain the mechanism for drawing the EMA^TM from a reel while simultaneously delivering the strand into the groove.  Also located on the head would be a sensor, cutting device and heating device.  In this design all mechanisms for delivering strand are located on the head.

 

 

Rotary Table
The rotary system depends on movements executed by a rotating motion.  This rotation would be coupled with a single direction linear movement allowing for travel to any coordinate designated by a control program.
 
One variation is based on a head suspended above a part.  The head would travel about a central axis as well as move horizontally.  It would also carry the mechanism to draw the strand from a reel and deliver the strand into the groove.  A sensor, cutting device and heating device would be located on this head.  All mechanisms for delivering strand are placed on the head.

The other variation rotates the part while the head operates suspended from a gantry.  While the part rotates, the head moves in the linear horizontal direction.  These two combined actions allow the head to enter any part of the groove.  The two motions, rotation and translation, would be concurrent for all the coordinates needed.

 

Single Head Design
In the single head design, the head is the device responsible for delivering the EMA^TM into the groove.  One structure has all the mechanisms necessary to place, soften, and cut the strand attached to a single head.  A placement device operates by knurled wheels gripping the strand and placing it in the groove.  A follower and curved guide would orient the strand in the groove by moving the strand from a vertical position to a horizontal position.  To prevent the EMA^TM from breaking during placement around corners, a heater would be used to soften the material and increase its pliability.  A sensor will detect the head’s return to the starting point.  This sensor also signals the cutting device to cut the strand to the proper length.

 

The Dual Head / Gripper Design
The dual head design contains the same components as the single head design.  However, the mechanisms are divided between two units, which creates simpler, less bulky systems.  The difficulty in the dual head concept lies in coordinating two moving structures.  One arm of the design maintains the head concept for the purpose of carrying the strand around the groove.  The other arm is nearly stationary and ensures the placement of EMA^TM into the groove.  This arm will hold the heating element along with a feeding mechanism for the strand.  Gripper arms for carrying the EMA^TM, the cutting device, and the sensor are all located on the head device.

 

 

Design Selection

Evaluation of the aforementioned design concepts is based on several criteria.  First the pros and cons of each variation are considered, as shown in Table 3.  Movements of the head versus the part, as well as the dual and single head concepts are compared below.  For the final design, the project team is selecting the x-y table design.  The x-y table is the most efficient and effective method for moving within the groove.  Another major factor in making this decision is that control of the x-y table is the simplest to integrate into a head design.  A gantry support will carry the head/heads.
 

	Pros	Cons
Move the Head	Smaller Machine Footprint	Strand Tracking
Move the Part	Increased Strand Accuracy	Larger Capacity Drives
Dual Head	Simplified Individual Heads	Increased Drive Controls
Single Head	One All-Purpose Head	Increased Complexity of Head

For this project, the head system is the most important technology to develop.  Either one or two heads could be used in conjunction with the x-y table.  There is no perceived advantage of creating a system with two independent heads.  Additionally, the complexity of coordinating the two heads is unfavorable when there is the option of using simpler head mechanisms.  A single head with all the placement, heating, cutting and sensing systems is more sensible for our purposes. Table 4 illustrates the design matrix used when deciding whether to use a single or dual head design.

A design matrix allows the different options for a design to be compared by various evaluative criteria.  The scale ranges from one (1) to ten (10).  A score of 1 reflects the option’s inability to meet design specifications; while a score of 10 means that the option is highly favorable in comparing it with the specifications.  Ideally, each criterion for a design would rank as a 10.  Since this is not usually the case, the total scores for each design option are compared to each other and to the ideal case.  The highest score indicates the option that most clearly meets the project’s needs.
 

Evaluative Criteria	Dual Head	Single Head	Nozzle	Ideal (Scale 1-10)
Speed	7	10	7	10
Cost	6	8	9	10
Robustness	9	8	9	10
Size	6	10	9	10
Safety	9	9	6	10
Maintainability	7	7	8	10
Precision	9	9	5	10
Total	53	61	53	70

The final design consists of the x-y table controlling a fixtured part.  A single head will be suspended above the part and contain all the placement systems.  Its controls allow for z-directional motion as well as 360-degree rotations.  Testing of the system is expected to prove that the head design has the capabilities to place strand.  Integration of the head with the x-y table is of lesser importance to the design scheme.  The head design must meet expectations for a prototype system and the control system must be suitable for a production type machine.  The provability is essential to the design.
 

Operation Description
1.  Initial Position:  The head is in the start position.  The part is moved along the assembly line below the head.  Strand is fed into the head.

2.  Alignment:  The z-axis motion control is used to lower the head in alignment with the grooved part.

3.  Strand Travel:  The strand is fed through the head using knurled wheels.  The strand passes through the nozzle, which is heated to soften the Ema^TM.

4.  The Guide:  After being heated, the strand then travels through the guide.  This guide is curved to ease the placement of the strand in the groove of the part.

5. Part Movement:  The XY table is pre-programmed with the shape of the current part.  At this point, the XY table begins to move according to the positions programmed and thus, the Ema^TM begins to fill the groove.

6. Strand Placement:  As the strand is placed the part moves according to the pre-programmed shape. Once the initial part of the strand is placed, the follower seats the strand in the groove.  This ensures that the entire strand is correctly placed on the bottom of the grooved part.

7. Proximity Sensor:  The proximity sensor locates the beginning of the strand placed to make sure that the requirements for the gap are met.
 
8. Cutting:  The strand is cut using pneumatic shears which are located on the head.
 
9. The head returns to the starting position.
 
10. The assembly line moves forward and the next part aligns for strand placement.
 

Control System

Following the completion of the mechanical design of the prototype, the electrical control system for the prototype was designed.  In order to simplify the design and construction of the machine, a control system will be purchased from Arrick Robotics.  This is a complete motion control solution that includes motors, drivers, power supplies, and cables. Software will need to be written for the specific application. The MD-2 Dual Stepper Motor System contains everything needed to accomplish motion control with an IBM style personal computer.  A block diagram of the system is shown in Figure 21 below. Up to 6 motors (3 MD-2 systems) can be connected to the computer.  The proximity sensor, heater, and cutter will all be controlled using this system.  A logic diagram can be found in Appendix E.  This diagram illustrates the logic that will control the strand placement.  It goes through the steps needed to lay strand in a part from start to finish.  This logic will be the basis for the software written to operate the control system.

A stepper motor moves in single steps that are usually .9 degrees each.  A coil inside the motor is energized and then works with the permanent magnets attached to the shaft.  Activating these coils on and off in sequence will cause the motor to either rotate forward or reverse.  A time delay between each step determines the motor speed.  Stepper motors can be moved to a position by actuating the requisite number of step pulses. Stepper motors have an advantage over servo-motors in that they can be used "open-loop" without the need for expensive encoders to check their position.  Stepper motors are more cost-effective than servo systems because of their simplified control and drive circuitry.  Maintenance is eliminated because there are no brushes to replace in a stepper motor. A stepper motor system cannot reach the speed of a servo motor system. The simplicity and ease of operation make them the preferred solution for many computerized motion control systems.   (www.arrick.com)

 

 Project Schedule

1. Finalize Part Specifications: Criteria for all off-the-shelf components were finalized and the parts to be ordered were selected.
 
2. Revise Drawings: Revise all detailed drawings from last semester.
 
3. Ordering Parts & Materials: Purchase orders must be completed and vendors contacted for the parts needed to build the prototype.
 
4. Proposal: Once our drawings and specifications were finalized and approved, a written proposal was compiled and the group presented their ideas to the department panel for review.
 
5. Build Custom Parts: Approved drawings will be submitted to a machinist for construction.
 
6. Construction of Test Fixture: A test fixture will be necessary to prove that the prototype device is functional and meets the design specifications set forth by Ashland Chemical.
 
7. Construction of Gear Box: The gearbox components will need to be assembled after they are machined.
 
8. Construction of Gantry: The gantry will be built as the support system for the head device.
 
9. Construction of Head Device: The head is the strand-laying device of the machine.  It will contain the heater, sensor, and cutting systems.
 
10. Progress Presentation: A presentation will be given to the department panel and Ashland Chemical to update them on the status of the project.
 
11. Control Panel Layout & Assembly:  All electronic components controlling the various systems of the device will be wired to and located on a main control panel.  This control panel will be adjacent to the constructed device.
 
12. Testing: All purchased components (heater, sensor, and drive mechanism) must be tested according to the developed test plans before being integrated with the main system. The computer program, which controls the x-y positioning table, will also need to be tested and debugged.
 
13. Testing of Prototype: After construction of the prototype, it must be tested and proven to work manually before being automated with the control system.
 
14. Troubleshooting: As testing is conducted, any problems or concerns that arise will need to be addressed and repaired.  This is to include any drawing revisions and/or reworking of parts.
 
15. Final Report: The last month of the semester will be used to conclude the second phase of the project. A report will be written to document the project activities for the construction and testing phase. This information will be presented to the department panel for their approval. A web page will also be created as a means of documenting information about the project.

 
 Budget and Resources

Specifications and prices were detailed for all required parts.  Once the final design was selected, the models to be used in our prototype were selected.  Appendix C contains a list of all custom parts that are included in the prototype design as well as the costs for materials and machining.  It also contains a listing of all off the shelf parts that are needed for the hardware and controls designs for the prototype.  Table 6 shows the summary of the Bill of Materials for the prototype including the costs of purchased component, custom parts, assembly cost, and all overhead expenses.  Assembly cost is derived from the cost of five junior engineers working ten hours per week for four weeks at $25 an hour.  This amounts to $17,500 per semester or a total of $35,000 for the year.  Assuming overhead costs to be 100% of labor costs, there will be additional costs equal to $17,500 each semester.  The total machining costs are estimated at $1,000 and the total estimated cost of all purchased and machined components with 15% overhead is $1764.68.  Therefore, the total estimated cost for the project is $72,765.
 

Item	Note	Cost
Labor	14 wks * 10 hrs * $25/hr * 5 members * 2 sem	$35,000.00
Overhead	100 % of Labor	$35,000.00
Machine Shop Time	$40/hr * 25 hours	$  1,000.00
Materials		$  1,764.68
Off the Shelf Parts	$                   1,291.50
Custom Parts	$                      243.00
Overhead (15% of Parts Cost)	$                      230.18
Total Budget		$72,764.68

 Financial Analysis

Economic Performance Predictions

As with any new industrial process an initial expenditure of resources is necessary to develop the fledgling concept into a mature and productive system.  These resources vary from the quantitative ones of time, money, space, and raw materials to the qualitative resources of knowledge, expertise, and management.  The intent of the Automatic Strand Layer is to cost effectively boost productivity and reliability of a process that is currently manually accomplished with an automatic machine.  This analysis will deal with the quantitative resources consumed by the machine in the development and consequent operational phases of the machine.  For the purposes of this study the design team will be considered as a subcontractor or development division of the sponsor company.  This results in the research and development cost being considered as a lump sum.  The minimum attractive rate of return (MARR) will be set at 20 percent and the inflation rate will be assumed to be steady at 3 percent for this study.  The development cost of the prototype is $72,765.  Production of the each fully automated machine can be accomplished at a cost of $8,320 each and selling price of $9,067.  This allows for a break-even point to be reached in five years according to the sales prediction table listed in Table 7 below.  Chart 1 illustrates the cash flows and net present value (NPV) predictions.  In regard to making a profit with this venture a selling price of $12,700 is suggested.  This is a 40 percent markup above the break sale price.   The NPV for this case is $3431,944 (Chart 2).
 

Year	Sales (# of units)
1	50
2	40
3	30
4	20
5	10

 

Machine Cost Performance
 
The economic performance of this machine is just as important as its technical capabilities.  Tables 8 and 9 compare the cost to produce 1000 parts using the automated machine and the current manual process.  For each case the time to produce one part is one minute.  This time can be reasonably achieved considering the strand feed rate of the machine and the skill of a manual laborer.  The machine’s costs include electricity, maintenance, depreciation, and that of a laborer tending the machine.  This tender’s labor would be necessary for one quarter of the operating time of the machine hence their labor is based on one quarter of the operating time.  They would provide such tasks as monitoring the operation, loading raw materials, and delivering finished pieces to the next stage of the assembly line.   The depreciation is straight line based on five-year machine life with an initial customer cost of $12,700.  The assembly worker is given a wage of $15 per hour and their overhead and benefits are estimated at 100% of their wage.  As it is apparent from the tables, the machine is capable of producing 1000 parts at approximately half the expense of a laborer only operation.
 

Machine Expenses
Parts	1000.0
Minutes per part	1.0
Operating time (mins)	1000.0
Operating time (hours)	16.7
$ Per Kwatt*hour	0.2
Kwatts required	4.6
Total Electricity Costs		11.50
Wear/repairs, per 1000parts		25.00
Depreciation Expense		4.83
Tender Expense		247.50
Total Production Cost		$ 289

The Automatic Strand Layer Project as it exists now consists of a pneumatic based cutter and height indexing platform, optical sensor linked to the cutter,  drive mechanism, heating system, and manually operated control panel.  Ashland Chemical viewed the prototype system and is pleased with the components.  The complete automation and integration of the components was not achieved due to budgetary issues.  However, Ashland Chemical views the overall project as beneficial because the project team was able to develop a head capable of feeding, directing, and delivering strand to a part groove.
 
In addition to the efforts of the student engineering group, a subcontracted professional automation corporation has been studying and developing a completely automatic turnkey system.  Ashland Chemical intends to share this prototype with them as well as the lessons learned during the project.  This consultant has experience in automating assembly lines, but since there are no current strand delivery systems available they are lacking a viable head design.  The mechanical and pneumatic components of the current head will be further developed and automated by the subcontractor, thus the custom student designed strand head will be meshed with commercially available automatic control hardware and software.
 
There are several areas of development that can be explored regarding the current head design.  Drive motors must be integrated as they are essential to automating the strand feed.  The double cylinder controlling platform height should be moved to provide a more stable design.  Additionally, it may be necessary to insulate the drive components from the heater and nozzle components.  This will allow the heat to transfer from the heater pipe down to the nozzle and not to compromise the bearings of the drive components.  This can be accomplished by using a ceramic bushing between the heater pipe and bottom plate, or by running cooling water through internal channels drilled in the bottom plate.  These design improvements and the subsequent automation should yield a commercially viable and automated strand delivery system.
 
 Conclusion

The EMABOND^TM process is an effective and efficient method for bonding thermoplastic parts.  One critical step in the process relies on the manual placement of the EMA^TM material in the mating groove.  The project team began by researching existing technology and incorporating it with their own innovative ideas to design the device.  A single head device containing all necessary components will be integrated with a commercial x-y table for the design.  The x-y table controls the motion of the fixtured part.  This design is the simplest and most functional of all the evaluated concepts.  Following the design phase will be the construction and testing of the prototype.

While the prototype is being constructed, individual components and assemblies were tested and debugged according to the test plans designed.  Due to budget constraints, Ashland decided not to automate our device at the present moment.  This is a functional requirement that engineers at Ashland will undertake with the assistance of a consultant specializing in automation.  The project goal to produce a working manual prototype of the head device has been successfully achieved.  Once the design is proven to function manually, the x-y table and control system will automate it to accomplish the tracing of the grooved part.

 Bibliography

“About Ashland” Ashland Chemical Homepage, 2000. http://www.ashchem.com.

Arrick, Roger, “MD-2 Dual Stepper Motor Control Systems” Arrick Robotics, 2000.  http://www.arrick.com.

Cengel, Yunus A.  Heat Transfer: A Practical Approach.  McGraw-Hill: Hightstown, New Jersey. 1998.

Delphion Intellectual Property Network, 2000. http://www.patents.ibm.com

 Acknowledgments

Ashland Chemical with special thanks to Michael Quinn, Marc Sacher, and Barry Sterrett

Professor Constantin Chassapis

George Wohlrab
 
 Appendix  Fall 2000 Semester Schedule

Methodology—Fall Semester

The project consists of several stages and is scheduled accordingly throughout the semester.  The stages consist of obtaining the project requirements from Ashland Chemical, researching existing technologies, developing conceptual designs, analyzing the design options and selecting the final design, and then generating detailed component and assembly drawings of the design.

The project description that follows refers to the tasks on the Fall Semester Gantt Chart located in Appendix B.  The numbers in parentheses correspond to the ordered tasks on the Gantt chart.

The goal of this project is to design a machine that automatically places EMA^TM strands into the grooves of injection molded plastic parts.  The first step in obtaining this goal is the problem definition (1).  The group began by consulting with Ashland Chemical about the project background.  During the initial consultation, we reviewed the process to be automated and its history.  Subsequent to learning about the problem, the group developed project specifications (3) with Ashland’s assistance.  We also performed background research on existing technologies (4).  Everything found related to this project was reviewed to assess whether any existing technologies could be adapted for our use.

Once the advisor and the group were knowledgeable about the project, weekly meetings and brainstorming session were held.  These resulted in several design alternatives (5).  All conceptual designs were drafted using Solid Works for comprehensibility.  Each of these designs was evaluated to determine their strengths and weaknesses.  There are several components that the group identified as ones that could be purchased from a vendor.  These components include the XY positioning table, heater, proximity sensor, and cutter.   Specifications for each of these components were written and potential vendors were researched (12).  Following another consultation with Ashland Chemical, a single-head design using an XY positioning table to control the movement of the plastic part was selected as the final design (19, 20).  At this meeting Ashland Chemical also identified the off-the-shelf components that they viewed as viable options.

Following these decisions, detailed engineering drawings (30) of the head were created using Solid Works and research on vendors was finalized.  At this point Ashland wishes to produce a full prototype of the single-head design.

 Project Schedule – Fall Semester

The project was outlined and scheduled using a Gantt chart.  The Gantt chart, which can be viewed in Appendix B, addresses each task for the project and also shows the time required for each task and the group member(s) responsible for completion of the tasks.

1.   Problem Definition:  The group learned about the background of the project, why it was being done, and the specifications for the device to be designed.
 
2.   Consultation with Ashland on Problem Definition:  In order to define the project and its specifications, the group needed to have an initial meeting with the engineers at Ashland.
 
3.  Develop Specifications:  In order to begin developing design concepts, the group must develop specifications to serve as guidelines during the design process.
 
4.  Background Research:  In addition to the information received from Ashland Chemical, a patent search was done to discover existing technologies similar to the project and to gain possible ideas for the design.
 
5.  Develop Design Alternatives:  The group held many brainstorming sessions in order to form ideas for the design of the automatic strand placer.
 
6.  Consultation with Ashland on Design Alternatives:  A meeting between the group, our advisor, and the engineers at Ashland was necessary so we could explain our many ideas for the machine’s design.
 
7.  Proposal:  Once our initial design alternatives were developed, a written proposal was compiled and the group presented their ideas to the department panel for review.
 
8.  Select Design Alternatives:  The group decided on the alternatives that they believed were the most feasible and best options to pursue further.
 
9.  Research Components:  The design for the head requires many components.  In order to decide which components we would buy from vendors and which we would make ourselves, some research needed to be done.  Specifications for each component (XY table, heater, sensor, motor controller, cutter, hardware/software interface) were developed and vendors were researched and contacted to find devices that suit our needs.
 
10.  Select Design:  After meeting with Ashland again, we were able to decide which design we would pursue as our final design.
 
11.  Development of Model:  A preliminary model was created using Solid Works so that we could clearly demonstrate our ideas during the upcoming meetings and presentations.
 
12.  Progress Report:  At the midterm we compiled a presentation for our department panel.  The purpose of this presentation was to explain our progress on the project thus far.  This presentation was also given to Ashland, so that they could inform us of any alterations they would like to make to the final design.
 
13.   Final Report:  The last month of the semester was used to conclude the design phase of the project.  A report was written to document the project activities for the design phase.  This information was also presented to the department panel for their approval.
 
14.  Assembly Component Drawings:  After Ashland approved the selected final design we modeled each component of the assembly using Solid Works.  These drawings will be used to machine the parts that cannot be purchased from vendors.
 
15.  Final Cost Estimate:  Once all the components were selected, the final cost estimate was calculated.  The results of this can be seen in Appendix C.