Design

Building from the previous interactions, the final design needed to address several factors. Tolerance within the different components needed to be adjusted to properly allow for smooth, repeatable motion. Using a spare piece of acrylic, several calibration holes were cut in order to properly size the hole for bolt we were using (figure 11). To achieve the desired smooth motion, we transitions away from conventional bolts to shoulder bolts which contain a shaft that is only partially threaded at the end.  In order to prevent jamming during motion, the design was restructured to approach but never reach the system's toggle point. To further increase the rope grip strength, the crank arm of the crank slider mechanism was elongated and the servo motor was moved to the side. This change allowed the servo to pull on the slider along it's line of action. This increased the closing force applied to the jaws by reducing the force the crank applied laterally to the slider. Also, to increase the gripping strength of the device, the jaws were offset from one another. This change forced the rope to serpentine through the jaws allowing for more friction to be applied to the rope. The offset jaws also reduce the load on the servo. Now when the rope in pulled, the resulting tension is opposed by the jaw and link assembly and not the the much weaker servo motor. A Solidworks model of the final design is shown in figure 12. In order to implement the offset jaw design acrylic spacers were used.

Figure 11 - Joint hole size calibration with shoulder blots.

Figure 12: Final model

Analysis

Using the same program as before, the kinematics of the mechanism were analyzed. This showed the location of the toggle point and at what angle the servo motor should initially be installed. The mechanical advantage of the system as a function of input angle is plotted below in figure 13.

Figure 13: Mechanical advantage vs. Input angle.

Mechatronics

Once the mechanics of the system were in place, the servo control program had to be built. Because our system needed to be attached to the body and somewhat mobile, we decided that a headless system (no monitor, keyboard, or mouse) would be most effective. Because of our previous experience, we decided to control our servo with a Raspberry Pi and a bread board. The bread board layout, shown below in figure 14, contains two switches. The first switch is used to change the desired position of the motor (open or closed), while the second signals the Pi to shutdown. This shutdown switch prevents memory card corruption that could be caused by abrupt power removal. Finally, an LED is added to signal that the servo control program is running. The servo is then attached to the Pi's PWM (pulse width modulation) pin to control the motor's position, detailed in figure 15. Next, the servo is powered by a six volt battery pack and grounded to the Pi.

Figure 14: Mechatronics circuit diagram.

Figure 15: Servo control with pulse width modulation.

The code pictured below, in figure 16, shows the servo control program. Every two seconds the Raspberry Pi checks to see the status of the switches and adjusts the position of the servo accordingly. If the shutdown switch is closed the computer cleans up all GPIO pins and begins to shutoff. Because of the lack of a real time clock on a Raspberry Pi, the position of the servo cannot be checked more frequently than once every two seconds. High frequency loops cause the sensitive timing of pulse widths, and thus the servo, to become unstable. Due to the nature of our project, quick changes in the jaw position are not necessary and a two second delay is acceptable. Finally, to make this a headless system, the control program must be executed when the computer boots up. This is done with a simple cron job that also defines the servo the open and closed position.

Figure 15: Raspberry Pi servo control code

Manufacturing

By building off of the lessons learned from the manufacturing of the previous prototypes, we were able to apply several design considerations into the final prototype. Carrying over the tolerance considerations of the laser cut acrylic, several in contact pieces were cut just oversized and then sanded to allow for a smooth and dynamic motion. Rather than using zip ties to attach the electric servo motor, we cut sandwiching acrylic pieces to link the motor to the crank slider. The use of shoulder bolts rather than standard threaded bolts kept the motion smooth and robust. Also, in order to lessen the effects of friction on the many links comprising the system, we used motor oil to grease the links and provide for a smooth operation. Finally, Velcro straps were added to the device to allow it to be attached to the inside of the user's hand. The links below contains a videos that demonstrates the device.

• Video 1
• Video 2

1Raspberry Pi Servo Guide: https://learn.adafruit.com/adafruits-raspberry-pi-lesson-8-using-a-servo-motor