Subsystem Build & Test
Table of Contents
Team Vision for Subsystem Level Build & Test PhaseWhat did our team actually accomplish during JUST this last week?
- Updated BOM to reflect our new design with more detail
- Updated problem tracking document with our implemented solutions and new problems from this week
- Contact Gates about different higher quality pulleys that are available
- Worked on Arduino side of software. Programmed based on a 12800 pulses per revolution motor input resolution.
- Tested the GUI and basic Arduino code with the motor. The motor registered the inputs and moved.
- Came up with new ideas to make the GUI more user friendly with more features.
- We plan to manufacture the power supply casing by Tuesday next week
- Modeled the calibration unit in Solidworks
- Marked angle brackets (taken as spare parts from machine shop) and cut to appropriate lengths for power supply enclosure
- Refined calculations for belt deflection. Some stretch will occur during movement (elongation equating to approximately 0.1 degree at maximum acceleration). However, once stopped, the belt will spring back to its desired location (the time between the motor stopping and the belt settling has yet to be determined, but it shouldn't be more than a few milliseconds). This could actually potentially act as the overshoot that occurs when human eyes perform a saccade.
- Updated overall CAD model to include both pulleys and belt
- Started presentation for Thursdays Review
- Set up a meeting with Dong for 9am Tuesday for further review of GUI
Updates on Belt and Pulley System
Last phase, the team progressed with a new system design which included a belt and pulley system to improve the output resolution without having to resort to a gearbox that would introduce error due to backlash. This concept was originally brought to our attention by Abe Amirana, our topic specialist from Teknic Motors.
Abe recommended using products from Gates, a manufacturer of high quality synchronous belts and sprockets. Abe has had some customers use motors from Teknic along with Gates products, and has seen these systems succeed. Another reason is the reputation Gates has with their quality of belt-drive systems. Lastly, we have decided to use Gates as our manufacturer of choice with our belt and pulley system components because the products are reasonably priced and fit within the project budget.
We have received a quote from Gates with the appropriate belt and sprockets for our assembly (see Bill of Materials below).
An additional part, recommended from the sales rep at Gates, is a belt idler. The idler will be used to maintain tension in the belt by pushing against the smooth side of the belt. Unfortunately, the belt idlers that Gates sells are much larger than what was recommended by their sales rep, but he did recommend we look at Stock Drive Products as a potential supplier for an idler that will suit our needs. The main specification that we are concerned with is the idler's outer diameter, which has to be small, but no smaller than 0.75 inch, according to the Gates sales rep.
Example of Belt Idler
We have also calculated the expected elongation to occur in the belt. The torque from the motor will effectively apply a tensile load of approximately 101 N to the belt as it pulls it around each of the sprockets at maximum acceleration. This load will result in an elongation in the belt that equates to approximately 0.1 degree in gaze position at the output. This elongation, however, will equilibrate as the motor comes to a stop at the desired position.
The elongation in the belt will create an inherent "overshoot" effect in the system, where the eye will travel slightly past its desired endpoint, but then snap back to its actual end position. This "overshoot" effect may perform similarly to how the human eye overshoots its target slightly when performing a saccade.
This "overshoot" effect will certainly be calculable. The team plans to meet with Dr. Kempski in the Mechanical Engineering department to discuss the system dynamics of our device during this upcoming phase. Upon our meeting with Dr. Kempski, we will be able to create a mathematical model of our system to determine exactly how much it will overshoot given a start and end point, along with any lag that the system will experience due to the stretch in the belt.
We have begun manufacturing some of our components. The components that we have machined thus far are as follows:
What Have We Manufactured?
Static Eye Clamp
The clamp for the static eye (shown below) has been machined to its desired state. We purchased two identical clamps inside which the OEMI-7 model eyes fit perfectly. For our static eye, we plan on using the threaded rod to adjust its height so it can stay level when the device is tilted to simulate diagonal eye motion.
In order to hold the acrylic together for the housing of the power supply, the angle brackets (shown below) had to be cut to the appropriate lengths, with chamfers cut into the ends at a 45-degree angle. These chamfers allow the angle brackets to hold all sides of the housing together without interference with each other.
What's Left to Manufacture?
These will be machined out of steel in order to keep high torque from deflecting or breaking the shafts due to high repetitions. We have acquired some cylindrical parts from machine shop, and these may be used to manufacture the shafts.
Our current design contains two separate plates of aluminum. These plates are planned to be purchased from Online Metals. Once they are received, the plates will be cut to the appropriate sizes with all necessary features added.
Dynamic Eye Clamp
Holes will have to be drilled from the front face through the rear face of the dynamic eye clamp to allow the calibration unit (see below) to be fastened to the front of the dynamic eye clamp.
Additional holes will have to be drilled into the flat sides of the clamp as well. These holes will be used with a pin to allow the dynamic eye to have an adjustable vertical gaze.
The yoke will be similar to the Alpha Team's yoke, but ours will be slightly wider in order for the thicker clamp to fit inside. The yoke allows the dynamic eye's gaze to be controlled horizontally by the motor, through the belt and pulley system, and also for a vertical component to be added to the gaze. The vertical gaze component will be manually adjustable.
The calibration unit (shown below) will be 3D printed, and used by placing it over the front of the dynamic model eye. The unit will have holes to match up with the through holes in the dynamic eye clamp so that the unit can be fastened to the clamp from the rear.
In the front surface of the calibration unit, we will bore out a space to embed a laser pointer diode. This diode will be able to show where the eye would be pointing, and would allow us to conduct our mechanical tests that determine the device's position resolution and accuracy, as well as its velocity and acceleration.