Tin & Tonic (Pre-Tinning Machine) (Team E)

Tyler Del Sesto Sebastian Käser Steven Rich Miller Sakmar

Mechatronic Design, 24-778|18-578 January 26, 2015

1

Table of Contents 1. Prefatory Information 2. Project Description 3. Design Requirements 4. Functional Architecture 5. Design Trade Studies 5.1. Feeding 5.1.1. Slide 5.1.2. Rotating Hopper 5.2. Sorting 5.2.1. Completing the Circuit 5.2.2. Slot Sorting 5.2.3. Light Reflection Sorting 5.3. Processing (Flux Application) 5.3.1. Actuated Syringe 5.3.2. Flux Pump 5.4. Processing (Wire Placement) 5.4.1. Feeding and Snipping 5.4.2. Pre-Cut Wire Loading 5.5. Processing (Part Placement) 5.5.1. Electromagnetic Placement 5.5.2. Tray Shift 5.6. Processing (Subsystem Interconnections) 5.6.1. Triple Head Assembly 5.6.2. Conveyor Processing 6. Cyberphysical Architecture 7. System Description 7.1. Overall Strategy 7.2. Feeding Slide 7.3. Circuit Sorting Method 7.4. Saw-Tip Spacer 7.5. Flux Injector 7.6. Wire Snipper 7.7. Part/Flux/Wire Applicator Robotic Arm 7.8. Completed Part Removal 8. Project Management 8.1. Schedule 8.2. Team Member Responsibilities 8.3. Budget 8.4. Risk Management 9. References

2

1. Prefatory Information (see title page…) 2. Project Description This project focuses on automating the pre-tinning process for saw tips manufactured by Kennametal. The current process requires: (1) manual sorting and placement of the saw tips, (2) manual application of Black Flux (Silver Brazing Flux), and (3) manual placement of two silver wire clippings. The machine described in this design proposal will automate these three tasks, enabling expedited preparation of pre-tinned saw tips in the Kennametal factory. This pre-tinning machine will also prevent repetitive motion strain injuries by removing human operators from the assembly process. Additionally, this machine will be easy to incorporate into the current product assembly line because it includes mechanisms to sort the raw materials before assembling the saw tips for oven treatment. 3. Design Requirements The device must completely process 20 parts (of at least two sizes) from initial pouring, through flux and wire application, and placement within five minutes. The flux must cover one-fourth to one-third of the blade’s top surface. The wires must be placed diagonally beside each other on the part to a maximum error of 5° in angle or 10% in length. The parts must end up, top-up, on a flat metal tray in 4 rows of 5, ½” from the edge and ¼” from each other, and be quickly removable from the machine. At least 19 of 20 parts must have correct wire placements, and the entire device must fit within a 2’x2’x2’ box. 4. Functional Architecture

3

5. Design Trade Studies In this section, the device’s subsystems illustrated above are described, and various methods for completing each task are detailed and compared. These comparisons enable educated design decisions to be made. 5.1. Feeding In this section, the saw tip pieces are poured into the feeding bin in random orientations, and the machine must place them in a horizontal position for further processing. 5.1.1. Slide In this design, the saw tips are poured on an angled slide. Half-way down the slide, after they have gained enough momentum, a ceiling higher than the shortest dimension of the block, but shorter than the other two dimensions, is added to knock the vertical blocks into a planar alignment. A conveyor belt at the bottom moves them through an opening in the side of the slide tall enough to allow horizontal tips to pass but short enough to knock vertical tips into horizontal alignment. Thus, the poured tips transition from random orientations to one of four planar horizontal orientations in one simple step. 5.1.2. Rotating Hopper In this design, the saw tips are poured into a hopper. At the bottom of the hopper is a rotating slat with a hole cut in it in the shape of a saw tip. Beneath one part of the slat is a hole leading to a short slide. As the slat turns, it picks up the pieces one by one, and drops them into the hole, ensuring once again that all the pieces are one of four flat orientations. This method is slower, and more prone to jamming than the slide described above but it provides a more regular output. It is also vulnerable small variations that can disrupt the feeding process. 5.2. Sorting Sorting takes the four orientation output from the feeder, and compresses it to two orientations, ensuring that the pieces maintain their topside up as they head toward the next steps for processing.

4

5.2.1. Completing the Circuit For this design, the saw tips slide through a point on the conveyor belt where they are in contact with the belt below and a fixed electrical sensor above. The electrical sensor consists of two prongs spaced a set distance part. They are to simultaneously contact only the top edge of the saw tooth piece, as the top edge is always longer than the bottom edge. This allows the sensor to differentiate between the long edge (top; circuit is completed) of the saw tips and the short edge (bottom, no electrical connection). This design has many benefits, including (1) simple and reliable sensing technique, (2) part spacing is maintained, and (3) limited mechanisms to reduce jamming potential. 5.2.2. Slot Sorting This method uses a conveyor belt and a slotted acrylic sheet to test the part configurations. First, the oriented parts coming from the feeding component are separated on the conveyor belt by a barrier that comes down regularly and only lets one saw tip through at regular intervals. The parts are then transported alongside an acrylic sheet with three different holes to match three of the different possible orientations. Each orientation follows a different path to get to the sorted conveyor belt. Another set of barriers are lowered onto the treadmill to stop the saw tips in front of each acrylic slot, allowing a solenoid to attempt to push the saw tip through the slot. If the part fits, then it slides through the to a specified mechanism. If the part does not fit, it stays on the conveyor belt and continues. Although capable of complete sorting, this method is quite complicated with a number of subsystems that are each vulnerable to failure. Furthermore, the mechanical slot sorting method is highly susceptible to jamming because of its need for high precision. 5.2.3. Light Reflection Sorting As the parts come from the feeding system, they are separated on the conveyor belt by a barrier that comes down regularly, only allowing one saw tip through at regular intervals. Once separated, a small light detector with a light source detects which side of the saw tip is facing up, using the reflected area of the saw tip to determine the side. If the top is already facing up, then the saw tip is pushed to the next stage. If the bottom of the part is facing up, then the part is flipped before moving to the next stage. Although this design is robust, it requires slow computation and depends on frequently inaccurate and challenging visual processing. 5

5.3. Processing (Flux Application) The flux application portion of processing concerns the painting of liquid flux onto the top of the saw tip, ensuring that it covers the proper proportion of the surface, before delivering the tips to wire applicator. 5.3.1. Actuated Syringe The syringe design resembles the current solution employed manually by Kennametal employees: the flux is carefully squirted by applying pressure via a continuous rotation servo connected to its plunger to the top of the plunger. The servo is attached to the plunger in a corkscrew fashion, similar to the screw on a chap stick container. Thus, as the servo turns, the plunger is pushed either up or down. This design has the advantages of high accuracy and airtight flux storage, which can extend the lifetime and reliability of the system. Additionally, it will provide a high range of accuracy for handling different-sized saw tooth pieces. 5.3.2. Flux Pump This design places a fluid pump at the bottom of the flux container. The pump intakes flux, outputting it through a 4/32” needle. One problem with this solution is the open container designed to hold the flux: the time scale of processing could be a problem because the flux separates over time. Another issue is the reliability and accuracy of DC pump for an unusual liquid such as flux. 5.4. Processing (Wire Placement) In this system, a pair of short silver wires are placed on the bead of flux centered on the saw tips in step three. 5.4.1. Feeding and Snipping In this design, two rolls of wires are fed into a wire holder. It consists of three vertical slats, with hinges at the top and special indentations at the bottom that hold each wire segments. These three segments are separated by servomotor actuators, the wire is fed inside, and the apparatus is closed. A blade slotted into the side of the set-up comes down and severs the wire segments from the spools. When the flux-covered saw tip is beneath the three slats, they open, allowing the two wires to fall in the correct orientation the small distance to the flux. By using spools of wire, this method eliminates the challenge of handling and organizing forty tiny wire segments. Thus, the design benefits from its reliability and efficiency. 5.4.2. Pre-Cut Wire Loading For this method, the precut wires are loaded into an angled vat. Like in the rotating hopper design for the saw tip loading, the bottom level of the vat, with two small wire6

shaped holes cut into it rotates. The tiny wire segments fall into these slits, which spin over another hole. They fall into two columns of stacked wires. A sliding door at the bottom controls the release of a single wire from each column onto the flux below. This design is particularly challenging because the hopper may not collect all the wires or because it may take a long time to do so. Additionally, it is difficult to release the wires one by one from the two stacks. Thus, inaccuracies and inefficiencies pose major challenges for implementing this method. 5.5. Processing (Part Placement) In this section, the saw tips are placed on the metal tray for delivery. Below are two strategies for this final step. 5.5.1. Electromagnetic Placement In this method, an electromagnetic hovers over the part and activates, picking it up. It then moves into position on the metal tray and releases the part. Since it touches the top of the part, it must be performed before the Pre-Tinning process. Although it requires a complicated apparatus, this design has high efficiency and precision which make it a good candidate for part placement. 5.5.2. Tray Shift In this design, a final conveyor belt is located over the terminal tray. When five parts come into position on the belt, the belt stops, and a pusher slides the parts onto the metal plate. The plate then shifts over one row to wait for the next set of five parts to be pushed onto its top. This design requires relatively intense coordination between various mechanical systems, precise timing and coding. Additionally, it can only handle parts arriving in regular intervals. Thus, its primary disadvantages are complexity, inefficiency and lack of robustness. 5.6. Processing (Subsystem Interconnection) This section describes the relationship between the flux application, wire placement, and part placement subsystems. 5.6.1. Triple Head Assembly In this design, all of the three processing parts are contained within a single mobile apparatus. This apparatus contains four copies of each processing component in parallel, and translates only on one axis to simultaneously process four parts. First, an 7

electromagnet lifts and places the part into position on the metal tray. Next, the head shifts until the flux applicator is positioned over the saw tip and places a spot of flux on the part. Finally, the wire applicator shifts over the part and releases two wires into the bead of flux. Since four parts are processed simultaneously, the device only needs to repeat this motion five times in order to complete the required twenty parts. The advantage of this method is efficiency: it can process all the saw tips very quickly and because there are fewer motions, error is less likely. However, tuning all these parts and constructing so many different heads may be time consuming. Nevertheless, the efficiency gains outweigh the other challenges. 5.6.2. Conveyor Processing In this design, the sorting system places all the parts one by one onto a conveyor belt. They pass beneath the flux applicator, where the belt pauses in order to apply a bead of flux. Next, they pass beneath the wire applicator where two wires are placed top while the belt is once again stationary. Finally, when they reach the appropriate position, they are pushed from the belt on to the plate. This method is quite slow since only one part is processed at once, and the belt must stop at each section. Transitions between subsystems are also difficult to facilitate. Additionally, with so many separate parts, and added complexity the probability for error is relatively high. 6. Cyberphysical Architecture

8

7. System Description 7.1. Overall Strategy The Pre-Tinning task can be separated in 3 different steps. First the tips must be sorted and orientated with the top facing up. Second flux and wire have to be applied on top of it and third the tips need to be placed on the tray. The sorting goes obviously first. This step is separated in two subsystems. The slide sorting system takes the tips as bulk good and outputs them on a conveyer belt. All tips with the “KERF” dimension perpendicular to the moving direction. They are then flipped and turned on the conveyer belt if necessary to have the top facing up by the circuit sorting system. Step two can be done on the conveyer belt or after placing the tips on the tray. We designed a system for either way to have a backup plan. A robotic arm, moving in X and Z direction grabs four tips on the conveyer belt and places them on the tray. The arm has a tri-functional head to complete step two as well. Different tools mounted on the same arm apply flux and wires on top of the tips once they are placed on the tray. If this turns out to be too hard to build or unreliable, the inversed option would be the backup plan. The same tools designed for the robotic arm may apply flux and wires when the tips are still on the conveyer belt. The finalized tips would then be sled from the conveyer belt onto the tray. 7.2. Feeding Slide The first step in the sorting was to, after the saw tips exit from the feeder in one of four orientations, making use of their asymmetrical geometry to passively rotate them. As soon as they come out of the feeder, the slide sideways down a slide, one by one. The slide must be the same width as the length of one of the parts so that a raised triangular section will induce a 180° turn in improperly shaped parts, while parts with a matching triangular cutout will pass over it unimpeded. This method is only able to eliminate one configuration, and upside-down parts still need to be flipped either before or after the slide. 7.3. Circuit Sorting Method Once the parts are lined up on the conveyor belt, a circuit sorting method is used the electrical conductivity of the parts to differentiate between the top (longer) and bottom (shorter) sides. To detect which parts are inverted, a smooth sensor is placed above the conveyor belt, leaving space for the saw tips to slide between the belt and the sensor. The sensor is designed to have two electrical contacts that are spaced wider than the length of the bottom

9

edge, but shorter than the top edge. As parts are pulled between the belt and the sensor, the sensor is pressed into the part using a passive spring to ensure a good electrical contact. If the circuit between the two electrical contacts is completed, then the part is known to be in the correct orientation. If a part passes under the sensor but does not complete the circuit, a side-over-side flipping mechanism is engaged that will flip the part to the correct orientation and replace it on the same place on the conveyor belt. This design has many benefits, including (1) simple and reliable sensing technique, (2) part spacing is maintained, and (3) limited mechanisms to reduce jamming potential. 7.4. Saw-Tip Spacer The tips coming out of the circuit sorting are not regularly spaced. The spacer ensures that the robotic arm may pick up four saw tooth pieces at a time, given that they are evenly spaced. This is necessary to put the saw tooth pieces on the tray properly. The saw tooth pieces arriving one by one on the conveyer belt are individually caught by the teeth of a second belt. This belt is perpendicular to the conveyer belt and driven by a stepper motor. It detects an arriving saw tooth piece by an inductive proximity sensor and goes forward by one tooth for every saw tooth piece. As soon as the four tooth has caught a fourth saw tooth piece, the robotic arm picks up the four saw tooth pieces all at once using the electromagnet.

7.5. Flux Injector Currently, the flux application system is one subcomponent of the Processing task. It is comprised of a continuous rotation servo motor, a 100mL syringe filled with flux, and a threaded plunger that fits the syringe. The syringe will be actuated by the radial movement of the servo motor, pushing the plunger down, and thus applying flux onto a variably-sized saw tooth piece. The design and motion of the threaded plunger is exactly akin to how a chap stick tube works; just as the bottom of the chap stick is turned to twist a screw-like post, which then pushes the chap stick plunger, which then pushes out the chap stick substance. This design will enable the flux application process to remain as close to the Kennametal employees’ process as possible. In addition, it may lead to easier flux replacement/renewal; the continuous rotation servo motor can retract the plunger back, taking in more flux. Or, the syringe could be modularly replaced as needed, with a new flux-filled syringe. Again, the plunger could be replaced back to its starting position whenever new flux is added. The microcontroller will attempt to keep track of plunger position by using a limit switch for the plunger starting position. As such, whenever the servo rotates, the rotation will be recorded and saved for future reference. The maximum rotation amount will have to be calibrated for the

10

syringe used, but will only need to be done once. The device described here closely resembles the patented mechanism [1]. 7.6. Wire Snipper Handling the wire pieces as bulk good seems to be difficult. Feeding the wire from a spool and cutting it just before applying it to the top of the tip simplifies the process. Two wires in parallel are pushed forward by two wheels trough the cutter. The wires are pushed into the gripping head, which will hold the cut wires and apply them to the tip. The wires are pushed till they hit the closed end of the gripper in order to control the length of the wire pieces. This allows us to use much cheaper DC motors to feed the wire instead of stepper motors that would be needed to control how much wire is fed otherwise. Then the cutter cuts the wire similar to a hand tool cable cutter. Tests need to be done to determine how much force is needed to be able to cut the wire. Based on these test the actuator for the cutter will be determined. There are two general possibilities, a solenoid with a lever or a DC motor with gears and a rack. If the force turns out to be too big for both methods, we would have to come back to the pre-cut wire loading. The wires will then be applied on top of the tips. A solenoid pushes down the head. This movement allows the jaws to be spread by a spring to release the wires. The head is rotated by 45 degrees to the orientation of the tip. This avoids a turning movement in order to apply the wires diagonally. For the robotic arm this system has to exist four times. If we are not allowed to feed from eight spools or other reasons prevent this possibility a Y movement have to be added to the arm (see 7.7) or to the wire snipper. 7.7. Part/Flux/Wire Applicator Robotic Arm The robotic arm move in X and Z direction. The head (1) of the arm is moving on two parallel axles in X direction. A stepper motor that drives two belts, adjacent to each axel and attached to the head, controls the X movement. A solenoid (2) mounted on the head may push down the different tools in Z direction. The Y direction is not needed as the arm has got all of its tools four times. Like this one line of the 5x4 tip matrix may be treated at once. This eliminates the need for a movement in Y direction.

1

11

X 2

1 Z

4

5

7

6

3

The head is tri-functional. First, electromagnets (3) pick up a series of four tips from the spacer (4) and put them onto the tray. Second, the flux capacitor tools (5) applies flux on the tips. Third, the head moves to the wire snipper (6) to pick up a series of wires with the gripping heads (7) and applies them onto the tips. 7.8. Completed Part Remover Once twenty parts are completed, the robotic arm assembly will return to the conveyor belt position and the system will wait for the user to remove the tray holding the twenty completed parts. After the finished tray is removed, the system will wait for the restart command for the user before continuing the automated task from the beginning. 8. Project Management 8.1. Schedule Week 3 (January 26 – February 1)

    

4 (February 2 – February 8)

    

5 (February 9 – February 15)

  

6 (February 16 – February 22)

  

Start website (Jan 30) Build first prototype of feeding mechanism Complete in-depth preliminary CAD of sorting and flux applicator Design skeleton code for system Prepare Mock-Up Demonstration (Feb 2) Unify design and devise subsystem transitions Construct flux pump prototype Test prototype for sorting sensing Complete in-depth CAD of wire cutters Sensors Lab (Feb 4) Start writing up code and organizing electronics Construct mobile rig for triple-tip applicator Construct mechanical skeleton Update Website (Feb 20) Complete feeding and flux pump mechanisms Integrate code, electronics with mechanisms

12

7 (February 23 – March 1)

   

8 (March 2 – March 8)

   

9 (March 9 – March 15) 10 (March 16 – March 22)



11 (March 23 – March 29)

  

12 (March 30 – April 5)

   

13 (April 6 – April 12)

 

14 (April 13 – April 19)

 

15 (April 20 – May 26) 16 (April 27 – May 3)



17 (May 4 – May 10)

  

18 (May 11)

 

  

 

Construct electromagnetic gripper Complete sorting mechanism Integrate sorting mechanisms code Basic linking of subsystems via conveyors Mid Semester presentation (Mar 2-4) Clean up subsystem linkages Work on triple-head code Construct wire-cutter prototype Spring Break/Catch-up Week Assemble first triple-head Complete triple head code Complete linkage of feeding and sorting Verify functionality of triple head Build two triple heads Attach triple head slider mechanism to machine Update Website (Apr 3) Build triple head Attach triple heads Clean up subsystem linkages Complete construction of machine Debugging and testing Debugging and Testing Prepare for Final Demo Final Systems Demo (Apr 22) Troubleshoot and/or beautify Start Final Report Public Presentation (May 6) Final Report (May 8) Peer Evaluation (May 9) Update Website (May 11) Clean up

8.2. Team Member Responsibilities Tyler Sebastian Steven Miller

Fabrication Helper Helper Leader

Assembly Co-Leader Co-Leader Helper

Electronics Helper Helper Leader

Programming Helper Helper Leader

13

Subsystem Responsibilities Tyler – Electric Circuit Sorting Sebastian – Part Spacer Mechanism, Robotic Arm Applicator Steven – Part Feeding and Sorting Miller – Flux Application, Microcontroller Use 8.3. Budget System Feeding/Sorting System

Flux System

Part Description

Qty.

Cost/Part ($)

Total Cost ($)

Timing Belt - 60T (217-3294) Timing Pulleys -18T (217-4100) 1/8 in. Acrylic Sheet 1/4 in. Acrylic Sheet

1 2 1 1

8.49 9.99 7.99 13.26

8.49 19.98 7.99 13/26

Syringe 100CC/ML Continuous Rotation Servos Parallax #900-00008 Stackable Stepper/Servo Controller System Power Supply: TekPower TP30SWII 30 Amp DC 13.8V 6061 Aluminum Angle Bracket 6061 Aluminum Rectangular Bar

4

7.2

28.8

4

14

56

2

19.95

39.9

1 1 1

99.95 13.43 22.66

99.95 13.43 22.66

Stepper Motor spacer belt Spacer belt Stepper motor arm Belt to drive arm Z solenoid Electromagnet DC Motor Timing Belt DC Motor Timing Belt Pulley Wire Cutter solenoid (ev. dc motor + gear) X guide axle 10 Z guide axle 8 Inductive sensor

1 1 1 2 1 4 1 2

14.95

14.95

19.95 14.95 16.58 6.99 17.46 15.21

19.95 29.90 16.58 27.96 17.46 30.42 16.58

1 2 4 1

16.58 8 3 13.95

Robotic Arm Applicator

16 12 13.95

14

8.4. Risk Management The biggest concern for in our project is sorting because we cannot know for sure how the tips will actually interact with our designed system. Thus, we will have to build prototypes of our different sorting systems in order to see which one works best. This will allow us to test all the aspects and potential failure reasons of these systems. That will be needed to finalize our design. The testing might also lead to the conclusion that one particular system does not behave like we expected. For this case we are prepared with our backup plans. We should not be afraid to use these plans while not giving up to soon on the original ones either. Error recognition is a task very little discussed. While verification at the end is only a coolness factor the Pre-Tinner needs to be able to detect if a tip gets stuck, a conveyer belt gets blocked or the robotic arm is unable to transition to the next state due to various reasons. A watchdog timer, which then puts our system in an error state, can detect latter. Counting the tips before putting them on the tray may be a solution to detect lost tips. Further the power consumption of the motors driving the belt can be monitored to detect a blocked belt. The problem with all these will be that we do have a limited amount of ports available for sensors detecting errors. As an experience of other projects time management will be a difficulty. Especially as this is a hardware based project. Delivery times for parts have to be taken in consideration. Deadlines have to be well planned and maintained, constant work on the project is essential. Last but not least the cost for prototyping is an issue. We will have to be careful about what to buy when and where from. First prototypes may be built from scrap found in the machine shop and parts we already own. 9. References [1] https://patentimages.storage.googleapis.com/EP2090188A1/imgf0010.png

15