Optical Rod Wrapping Machine
This project was intended for the UH Manoa Physics lab to increase the efficiency of lab testing. The task was to build a machine that would wrap optical rods with Teflon tape uniformly and efficiently. This project offered the opportunity to learn about the physical development of machines, working with stepper motors. Eccentric nuts and various aspects of physics such as torque loss over a shaft and adjustable friction knobs for the tension of the Teflon tape. Furthermore, I learned about programming GUI's for the Raspberry Pi from another team member.
To make this project less monotonous please take a look at the first picture of this project publication and go to the section labeled to see what I learned.
Section 0: Initial Design
The goal of the machine was the be able to wrap rectangular prism rods with Teflon efficiently and consistently. The initial design consisted of an aluminum extrusion frame with two motors. One motor at one side of the rod and a free rotating bearing at the other side. This rotated the rod along its longest axis. We intended to use rubber feet to keep the rod in place without interfering with edge to edge coverage of the rod. The other motor moved the carriage holding the Teflon tape along the length of the rod to wrap the rod evenly. Though the general design stayed, we learned and made changes to ensure the success of the projects as you will read about in the following sections.
Section 1: The Frame
The machine's frame was made using 20 x 20 aluminum extrusion with t-slots. The use of this material has many advantages. It is readily available and easy to work with. A bandsaw or Sawzall can be used to cut the pieces to length. The keyways allow us to connect the pieces using simple 3d printed parts and to mount different pieces to the frame at any location. The flexibility of use is one of the main reasons it was chosen versus the fabrication of a frame.
The six pieces are held together using 3d printed angle brackets. One of the challenges was that if we designed the angle bracket exactly symmetrical, then the T-slotted framing fastener would not fit once the first one was installed. Due to this, one of the holes had to be offset to avoid hitting the head of the screw. (Figure 1.1)
Section 2: Stepper Motors
Before taking on this project, the team and I were already familiar with stepper motors from 3-d printing. However, we had never programmed or used a CNC shield to control them. Spec sheets helped a lot when developing the code. The tooth count for stepper motors was crucial as it controls the degree of rotation per step. Given the constraints of the project, we chose off-the-shelf NEMA 17 stepper motors with an Arduino and A4988 Motor driver. This motor has 200 steps which means a 1.8-degree step angle. This was precise enough as we did not need to move the motor to less than 5 mm. Particularly this was more involved in the roller carriage aspect than the optical rod rotation. A plethora of calculations were done to move the carriage at the right speed to have the Teflon overlap on the last rotation of Teflon.
The stepper motor mount that rotates the rod also had challenges in its practical use versus the initial design. Should the machine be remade, the mount would have mounting screws outside of the frame of the motor mount to adjust it without having to take off the CNC motor. This would lessen the length of the rod that could be wrapped but would increase the flexibility of use. (Figure 1.2)
Section 3: Roller Carriage + Belt Drive
Having a background in 3-D printing also helped with implementing the initial design. Most 3-D printer heads use a v slot wheel to roll along an extrusion and are belt driven. This mechanism is what we used for the roller carriage that moves the Teflon along the length of the rod. The carriage was the second most intricate design on this project as it had three duties. The carriage rolled on the extrusion while holding the belt and held the Teflon with adjustable friction to ensure the Teflon had enough back pressure for a tight wrap. The most challenging part was mounting the v slot wheels on the carriage. The team and I overlooked the use of eccentric nuts (Figure 1.3) on carriage heads for 3d printers. The eccentric nut allows for tightening and loosening of the carriage because of wear and tear on the wheels. If the machine were to be remade, that solution would be implemented. Another wear and tear adjustment device used is the belt tensioner on the left end of the extrusion. This allowed us to adjust the belt tension in case it came loose instead of having a static tension.
The end stop used for calibration was also inspired by 3-D printing To verify where the carriage is, the first step of every wrap was to move the carriage to the end stop at the far right. This ensured the Teflon was aligned with the right end of the rod.
Section 4: Arduino + CNC Shield
The setup of the CNC motors was straightforward as there was a lot of documentation concerning the NEMA 17 and CNC shield. A plethora of calculations was involved in getting the rod the wrap right, but after the reference data was created, our team was able to create multiple programs for various lengths of rod.
Section 5: Torque Loss
The initial design only had one motor on the right side of the rod to provide the torque, and the other side would be on a free rotating bearing. At this point, the team and I remembered why cars have engine and wheel horsepower. Torque loss across a shaft caused the rod to slip off the mount at the desired speed. Furthermore, when we sped the process up, the rod would get wrapped on the motor side but fall off the mount on the bearing side. This caused us to have to add a second motor to alleviate the stresses the rod was experiencing.
Section 6: Tension Via Friction
One of the reasons why the rod would fall off is the tension we had to apply to the Teflon tape to cause a slight stretch that would wrap the rod tightly. Because the rolls of Teflon would have to be changed frequently, we had to design a system where we could adjust the friction on the Teflon roll to achieve the desired wrap. A printed washer was the initial design where the washer would rub against 10% of the inner circle of the Teflon roll. We discovered that the consistency of the Teflon rolls is not controlled very well, and most rolls were warped. This caused a design change where we took advantage of the surface area around the roll and printed a carriage with walls that would provide consistent friction instead of dynamic.
Section 7: CAD Design & 3-D printing
The most involved 3-D printed part was the housing for the electronics. We needed to hold the Raspberry Pi in the box and access the ports as well as the Arduino CNC motor shield and monitor. Due to the time it would take to print this piece, all aspects of the design had to be nearly perfect, so a reprint would not be necessary. To achieve success on the first try, the Rasberry Pi mount and Arduino mount were brought in from Thingiverse. These models were joined in Fusion into the box and verified to allow enough space. Cutouts were made for the wiring in the box. An improvement can be made by filleting the outer edges of the box to make for a better look and feel.
Because the box was enclosed, heat dissipation became a concern. Specifically, the CNC shield and Rasberry Pi produce a lot of heat, and a fan and vents were added to prevent overheating and improve the machine's longevity.
The end result was not as perfect as expected. The supports used when 3d printing an overhang was not easy to remove. If printed again and in my own designs, minimizing the use of supports is crucial for good prints. This ensures that the print comes out cleaner and has less opportunity for failure when printing. To do this tapers and integrated supports are great options as they do not need to be removed after printing.
Section 8: Raspberry Pi and Controls
Though this machine can be used as a standalone device for one length and size of optical rod with an Arduino alone, it was important that values of length and girth were adjustable directly from the machine. This is to ensure that one user’s computer does not have sole control over changing the code and keeps the machine dynamic for other tasks. For this reason a Rasberry Pi with a touchscreen was added. Though this did not have direct control over the Arduino it was used to upload new code when necessary to change variables. This code was premade using what was learned from the other sections of the project to ensure consistent performance over different types of rods. Furthermore the premade code could be edited with the calculations on the Rasberry Pi continuing to allow for ease of use and a community based use case.
To make this project less monotonous please take a look at the first picture of this project publication and go to the section labeled to see what I learned.
Section 0: Initial Design
The goal of the machine was the be able to wrap rectangular prism rods with Teflon efficiently and consistently. The initial design consisted of an aluminum extrusion frame with two motors. One motor at one side of the rod and a free rotating bearing at the other side. This rotated the rod along its longest axis. We intended to use rubber feet to keep the rod in place without interfering with edge to edge coverage of the rod. The other motor moved the carriage holding the Teflon tape along the length of the rod to wrap the rod evenly. Though the general design stayed, we learned and made changes to ensure the success of the projects as you will read about in the following sections.
Section 1: The Frame
The machine's frame was made using 20 x 20 aluminum extrusion with t-slots. The use of this material has many advantages. It is readily available and easy to work with. A bandsaw or Sawzall can be used to cut the pieces to length. The keyways allow us to connect the pieces using simple 3d printed parts and to mount different pieces to the frame at any location. The flexibility of use is one of the main reasons it was chosen versus the fabrication of a frame.
The six pieces are held together using 3d printed angle brackets. One of the challenges was that if we designed the angle bracket exactly symmetrical, then the T-slotted framing fastener would not fit once the first one was installed. Due to this, one of the holes had to be offset to avoid hitting the head of the screw. (Figure 1.1)
Section 2: Stepper Motors
Before taking on this project, the team and I were already familiar with stepper motors from 3-d printing. However, we had never programmed or used a CNC shield to control them. Spec sheets helped a lot when developing the code. The tooth count for stepper motors was crucial as it controls the degree of rotation per step. Given the constraints of the project, we chose off-the-shelf NEMA 17 stepper motors with an Arduino and A4988 Motor driver. This motor has 200 steps which means a 1.8-degree step angle. This was precise enough as we did not need to move the motor to less than 5 mm. Particularly this was more involved in the roller carriage aspect than the optical rod rotation. A plethora of calculations were done to move the carriage at the right speed to have the Teflon overlap on the last rotation of Teflon.
The stepper motor mount that rotates the rod also had challenges in its practical use versus the initial design. Should the machine be remade, the mount would have mounting screws outside of the frame of the motor mount to adjust it without having to take off the CNC motor. This would lessen the length of the rod that could be wrapped but would increase the flexibility of use. (Figure 1.2)
Section 3: Roller Carriage + Belt Drive
Having a background in 3-D printing also helped with implementing the initial design. Most 3-D printer heads use a v slot wheel to roll along an extrusion and are belt driven. This mechanism is what we used for the roller carriage that moves the Teflon along the length of the rod. The carriage was the second most intricate design on this project as it had three duties. The carriage rolled on the extrusion while holding the belt and held the Teflon with adjustable friction to ensure the Teflon had enough back pressure for a tight wrap. The most challenging part was mounting the v slot wheels on the carriage. The team and I overlooked the use of eccentric nuts (Figure 1.3) on carriage heads for 3d printers. The eccentric nut allows for tightening and loosening of the carriage because of wear and tear on the wheels. If the machine were to be remade, that solution would be implemented. Another wear and tear adjustment device used is the belt tensioner on the left end of the extrusion. This allowed us to adjust the belt tension in case it came loose instead of having a static tension.
The end stop used for calibration was also inspired by 3-D printing To verify where the carriage is, the first step of every wrap was to move the carriage to the end stop at the far right. This ensured the Teflon was aligned with the right end of the rod.
Section 4: Arduino + CNC Shield
The setup of the CNC motors was straightforward as there was a lot of documentation concerning the NEMA 17 and CNC shield. A plethora of calculations was involved in getting the rod the wrap right, but after the reference data was created, our team was able to create multiple programs for various lengths of rod.
Section 5: Torque Loss
The initial design only had one motor on the right side of the rod to provide the torque, and the other side would be on a free rotating bearing. At this point, the team and I remembered why cars have engine and wheel horsepower. Torque loss across a shaft caused the rod to slip off the mount at the desired speed. Furthermore, when we sped the process up, the rod would get wrapped on the motor side but fall off the mount on the bearing side. This caused us to have to add a second motor to alleviate the stresses the rod was experiencing.
Section 6: Tension Via Friction
One of the reasons why the rod would fall off is the tension we had to apply to the Teflon tape to cause a slight stretch that would wrap the rod tightly. Because the rolls of Teflon would have to be changed frequently, we had to design a system where we could adjust the friction on the Teflon roll to achieve the desired wrap. A printed washer was the initial design where the washer would rub against 10% of the inner circle of the Teflon roll. We discovered that the consistency of the Teflon rolls is not controlled very well, and most rolls were warped. This caused a design change where we took advantage of the surface area around the roll and printed a carriage with walls that would provide consistent friction instead of dynamic.
Section 7: CAD Design & 3-D printing
The most involved 3-D printed part was the housing for the electronics. We needed to hold the Raspberry Pi in the box and access the ports as well as the Arduino CNC motor shield and monitor. Due to the time it would take to print this piece, all aspects of the design had to be nearly perfect, so a reprint would not be necessary. To achieve success on the first try, the Rasberry Pi mount and Arduino mount were brought in from Thingiverse. These models were joined in Fusion into the box and verified to allow enough space. Cutouts were made for the wiring in the box. An improvement can be made by filleting the outer edges of the box to make for a better look and feel.
Because the box was enclosed, heat dissipation became a concern. Specifically, the CNC shield and Rasberry Pi produce a lot of heat, and a fan and vents were added to prevent overheating and improve the machine's longevity.
The end result was not as perfect as expected. The supports used when 3d printing an overhang was not easy to remove. If printed again and in my own designs, minimizing the use of supports is crucial for good prints. This ensures that the print comes out cleaner and has less opportunity for failure when printing. To do this tapers and integrated supports are great options as they do not need to be removed after printing.
Section 8: Raspberry Pi and Controls
Though this machine can be used as a standalone device for one length and size of optical rod with an Arduino alone, it was important that values of length and girth were adjustable directly from the machine. This is to ensure that one user’s computer does not have sole control over changing the code and keeps the machine dynamic for other tasks. For this reason a Rasberry Pi with a touchscreen was added. Though this did not have direct control over the Arduino it was used to upload new code when necessary to change variables. This code was premade using what was learned from the other sections of the project to ensure consistent performance over different types of rods. Furthermore the premade code could be edited with the calculations on the Rasberry Pi continuing to allow for ease of use and a community based use case.
2 Likers
Comments