1 1. Problem Definition Supporting Document 1.1 Annotated Bibliography There have been numerous iterations of the ME 3281 lab kit spanning the last decade. While none of them completely embody the ideal lab kit, each iteration has evolved and improved upon the last. It is very important to carefully and thoroughly research these previous lab kits so past mistakes can be avoided and efforts are not duplicated. This research mainly focused on the mechanical components of the lab kit, namely the motor, mass, springs, and position sensors. Much of the research effort involved narrowing down the abundance of resources compiled by previous senior project groups. The previous groups also documented past work so there are many sources and documents to review and eliminate when the information is redundant. Many of these sources are academic publications and product descriptions. [1] Durfee, Li and Waletzko, June, 2004 “Take-home Lab Kits for System Dynamics and ControlsCourses”. Proceedings of the 2004 American Controls Conference. This document provides information about the design of the first generation of the take-home kit. It is a 4 th order system composed of multiple springs, masses, and a damper. Knowing the history of this project has provided guidance in the design of the 2013 iteration. This information has allowed the team to learn from mistakes, and benefit from strengths. [2] Durfee, Li and Waletzko, June, 2005 “At-home System and Controls Laboratories.” Proceedings of the 2005 American Society of Engineering Educators Conference. This document covers more information regarding the 1 st generation lab kit. It provides further insight into the motivation and implementation of the original iteration of this project. This source was used as background information for 2013 take-home kit design. [3] Waletzko, D. December, 2005 “Distributed Laboratory Modules for System Dynamics and Controls Courses”, A Plan A Master’s Thesis, University of Minnesota, Institute of Technology. Waletzko’s thesis was another helpful source providing design details about the 2 nd generation kit. It is a comprehensive documentation describing its implementation. The rotary system used in this kit is used as the basis for the 2013 iteration. Gaining insight from as many different designs and revisions of this kit provided guidance for new kit design. [4] ME4054 Work, Available: http://www.me.umn.edu/dlab/ME4054work.html [Accessed: March 3, 2013] This site provides information about another kit design, done by senior design students in the Spring of 2003. It details each component of the kit. This was another design which contributed to lessons learned and decisions for the new kit.
43
Embed
1. Problem Definition Supporting Document · 1. Problem Definition Supporting Document ... Close, Charles, Dean Frederick, and Jonathan Newell. Modeling and Analysis of Dynamic Systems.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
1. Problem Definition Supporting Document
1.1 Annotated Bibliography
There have been numerous iterations of the ME 3281 lab kit spanning the last decade. While none of
them completely embody the ideal lab kit, each iteration has evolved and improved upon the last. It is
very important to carefully and thoroughly research these previous lab kits so past mistakes can be
avoided and efforts are not duplicated. This research mainly focused on the mechanical components
of the lab kit, namely the motor, mass, springs, and position sensors.
Much of the research effort involved narrowing down the abundance of resources compiled by
previous senior project groups. The previous groups also documented past work so there are many
sources and documents to review and eliminate when the information is redundant. Many of these
sources are academic publications and product descriptions.
[1] Durfee, Li and Waletzko, June, 2004 “Take-home Lab Kits for System Dynamics and
ControlsCourses”. Proceedings of the 2004 American Controls Conference.
This document provides information about the design of the first generation of the take-home kit. It is
a 4th order system composed of multiple springs, masses, and a damper.
Knowing the history of this project has provided guidance in the design of the 2013 iteration. This
information has allowed the team to learn from mistakes, and benefit from strengths.
[2] Durfee, Li and Waletzko, June, 2005 “At-home System and Controls Laboratories.”
Proceedings of the 2005 American Society of Engineering Educators Conference.
This document covers more information regarding the 1st generation lab kit. It provides further insight
into the motivation and implementation of the original iteration of this project.
This source was used as background information for 2013 take-home kit design.
[3] Waletzko, D. December, 2005 “Distributed Laboratory Modules for System Dynamics and
Controls Courses”, A Plan A Master’s Thesis, University of Minnesota, Institute of
Technology.
Waletzko’s thesis was another helpful source providing design details about the 2nd generation kit. It
is a comprehensive documentation describing its implementation. The rotary system used in this kit is
used as the basis for the 2013 iteration.
Gaining insight from as many different designs and revisions of this kit provided guidance for new kit
design.
[4] ME4054 Work, Available:
http://www.me.umn.edu/dlab/ME4054work.html [Accessed: March 3, 2013]
This site provides information about another kit design, done by senior design students in the Spring
of 2003. It details each component of the kit.
This was another design which contributed to lessons learned and decisions for the new kit.
2
[5] Jouaneh, M. and Palm, W., June, 2009 "System Dynamics Take-Home Laboratory Kits", In
Proceeding of the 2009 ASEE Annual Conference, Austin, TX.
This document describes the take-home lab kit developed by Rhode Island University. It contains 4
different kinds of lab manuals and gives information about kit design and use.
There is no known patent surrounding a kit of this type, so the design team was able to use the ideas
implemented in other kits for the benefit of the 2013 design.
[6] Quanser Inc, Available:
http://www.quanser.com/flippers/Rotary/2012/ [Accessed: March 3, 2013]
Quanser Inc produces a commercial lab kit that teaches system dynamics and controls concepts.
This source provided some ideas that could be used in the take-home kit design, such as the concept
of interchangeable components that may be used so students can understand their effects on system
performance.
[7] Educational Control Product (ECP), “Industrial Plant Emulator”, Available:
http://www.ecpsystems.com/controls_emulator.htm. [Accessed: April 5, 2009].
Educational control product is a kit used in industry or an educational setting. It gives details on
interchangeable components as well as basic concepts of spring and mass configurations to our kit.
This is yet another example of a source which was used to refine the design of the 2013 take-home
kit.
[8] Distributed Laboratories, Available: http://www.me.umn.edu/dlab/ [Accessed: March 6,
2013]
Reference [8] is a central index of the efforts that have been made towards designing and providing a
take-home ‘distributed’ lab kit to students. It indexes the people involved, initial proposal, relevant
progress and publications, as well as various other software and files for download. This is the place
to begin when understanding the background of this subject.
The distributed laboraties page was used to understand the background and various iterations that the
take-home kit has undergone. The decisions made in the past, as well as lessons learned, are a helpful
1.) Do you feel that ME 3281 would benefit from a take home kit? 73.81%
2.) Do you have an Arduino Microcontroller? 86.90%
3.) Would you be willing to pay around $50 for such a kit in addition
to your textbook? 39.29%
4.) From the following list, select all topics that you feel could
benefit from an interactive demonstration.
a) System Modeling 67.86%
b) Step Response 59.52%
c) Frequency Response 71.43%
d) PID Control 84.52%
Table 1-1: User Need Survey Results
The intent of the first question was to determine if there was a need for the product. From the
survey results, it can be seen that 73.8% of students felt that the class could have benefited from a lab
component, verifying that the market exists.
The second question was used to validate our intent to use the Arduino Microcontroller from
the ME 2011 class as the lab kit controller. As can be seen from Table 1-1, 86.9% of the class has an
Arduino of some form.
The previous kit models have all cost around $50 a piece. The survey results from question 3
show that only 39.29% of students are willing to pay such a cost. This number increases slightly to
51.6% when calculated using only the students who felt the lab kit would be beneficial. From these
results, it is clear that low cost is a very important design criterion.
The last question was used to determine what topics would be most beneficial to cover in the
lab. From the results it can be seen that PID control is the most requested topic, followed by
Frequency Response, System Modeling, and Step Response. These results will be used to design the
lab curriculum, focusing more on areas that the students felt were confusing.
These survey results were combined with needs determined from our own experiences from
ME 3281 to form a list of user needs. The importance rankings were based on the survey response
data, as well as our own intuition from previous experience in the class. This list is provided in Table
1-2.
6
# Customer Need Importance Source
1 Will effectively teach ME3281
Curriculum 5 Survey
2 Will be reasonably priced 4 Survey
3 Will integrate Arduino microcontroller 3 Survey
4 Will be easy to construct 2 Team List
5 Will be reasonably sized 2 Team List
7 Will be reliable and rugged 4 Team List
9 Will collect position data accurately 4 Team List
10 Will operate safely 5 Team List
11 Software is intuitive and easy to set up 4 Team List
12 Will utilize Matlab user interface 3 Team List
13 Software incorporates platform
independence 4 Team List
14 Will determine velocity accurately 4 Team List
Table 1-2: Ranked List of User Needs
1.4 Concept Alternatives
With regard to the physical makeup of the take-home kit, there are 4 subsystems that require
concept consideration and selection. Additionally, there are also multiple high level system
configurations to be considered. Lastly, a robust software package must be designed. The
TAKEHOME team has considered various proposals of each system and has made a selection based
on objective criteria available for each. Past experiences and lessons learned have weighed heavily
into this process, and will be leveraged for the most effective and highest quality overall outcome.
Type of System
In the most general sense, the system to be replicated with the ME3281 take-home kit is a
traditional mass-spring-damper system. This is a second-order system with two energy storing
elements; the mass and the spring. Within this, two types of systems can be created: linear and
rotational.
A linear system was created as a part of 2004 work on the take-home kit project. This was
modeled as a ‘quarter-car’ linear translational model, with 2 springs, 2 masses, and a damper. The
system representation is shown in Figure 1-1. The top mass and spring can be removed to convert it to
second order. This system requires alignment and control of the directional movement of the mass,
and a method of converting the rotational motion of the motor into linear translation, which introduce
complexities.
7
Figure 1-1: Quarter car model translational system
A rotational system has been the most common method to create a simple mass-spring-
damper system, both in take-home kit project work, as well as commercially available systems. This
system is easier to produce, as the output shaft of the motor can be directly coupled to the system to
drive input. A system representation is shown below in Figure 1-2. However, there are challenges in
creating a linear damping component, as well as spring configuration, which will be discussed in
subsequent subcomponent sections.
Figure 1-2: Rotational system with motor torque input
Spring Alternatives
The spring selection has been an area of trouble for prior revisions of the take-home kit. A
rubber band was wrapped around the mass and a stationary post to provide a spring force about the
central shaft. See Figure 1-3 for illustration from the 2009 4054 project report. This design was
advised against by project advisor Professor Durfee. The rubber bands are subject to significant
degradation, causing spring constant changes, as well as breakage. This potentially leaves the student
without a spring element.
8
Figure 1-3: Wrapped rubber band spring configuration
The next spring concept considered was that of a metal torsional spring. This type of spring
is specifically designed for rotational systems, provides a very linear spring force, and could be
packaged easily in the take home kit system. However, this type of spring is very difficult to find in
spring rates low enough to be suitable for this kit, and is designed to only be operated in one direction
of rotation. See Figure 1-4 for illustration of spring configuration
Figure 1-4: Torsion Spring, pulley, motor shaft configuration
In order to keep the kit simple and affordable, it is desirable to attempt to incorporate more
traditional coiled linear springs into the rotational system. An initial idea was to use two springs that
are mounted at points 180 degrees apart on the pulley, and to fixed positions on the lab kit base. This
configuration is shown in Figure 1-5. This concept has the benefit of simple, compact construction.
However, there is concern over the spring coils ‘binding’ as they come into contact with the pulley,
resulting in severe rotational limits. There is also a lack of perfect linearity as one spring relaxes by a
different distance as the opposite spring extends during rotation.
To resolve the drawbacks of the previous concept while still integrating linear coil springs,
another design is considered. This concept is achieved by using a single mount point on the pulley,
and cable to alternatingly pull springs from a fixed mount point on the unit baseF. This design allows
the use of very cheap, small and readily available coil springs. Rotation is limited around 90
9
degrees, and requires a mount point some distance out from the pulley. Exact mount distance depends
on the relaxed spring length. See Figure 1-6 for illustration of this concept.
Figure 1-5 (left): 180 degree separated spring configuration, mounting points shown in red Figure 1-6 (right): Shared mount configuration with cable, mounting points shown in red
Damper Alternatives
Damper selection and incorporation has been a recurring problem in past 3281 take-home kit
revisions. Various ideas have been considered, but ultimately thrown out for various reasons, in
favor of using internal system friction as the damping agent. This is a less than ideal solution, so
damper selection is once again revisited. Friction is a less than ideal damping method, as the damping
term is linearly dependent on the angular velocity of the system, which introduces a non-linearity and
causes non-ideal system behavior. Taking into consideration the fact that this kit is meant to closely
approximate an ideal system for the benefits of student learning, friction must not be allowed as the
only damping component.
The first type of damping component considered is one of the many commercially available
offerings. The principle behind the operation of such devices is the internal use of a rotating disc
submerged in viscous fluid, along with various valves and vanes to direct fluid flow. This has the
end result of very linear damping with respect to rotational velocity. However, these components
suffer from a few drawbacks. The low availability of dampers small enough for this application, as
well as a high cost (around $3.50 each), makes them less than ideal. Additionally, these units are rated
for very low rotational speeds; 10 cycles per minute. See Figure 1-7, below, for a patent schematic
of a rotational damper.
10
Figure 1-7: Patent schematic of a common type of rotational damper
Using similar general principles behind the viscous rotational damper, another concept is
proposed that replicates a similar behavior in a simple and inexpensive way. A small fan blade
could be incorporated that sits inside a cup of water or other viscous fluid. When spun by the motor,
this blade will produce a damping effect. This damping force is due to a drag force generated by the
blade against the liquid, quantified by Equation 1-1.
𝑓𝑑 = 12⁄ 𝜌𝑈2𝐶𝑑𝐴
Equation 1-1: Drag Force
As shown, this force varies quadratically with velocity, which again represents a non-
linearity. However, the effect of this term can be reduced by using a fluid significantly more dense
than air (water, oil) and by limiting velocity ‘U’ to very small values. This will replicate a linear
damping factor over the limited breadth of operation of the take-home kit. See Figure 1-8 for a rough
physical layout of this concept. One noticeable impact of this design is the necessity to mount the
motor with the shaft facing downwards.
Figure 1-8: Proposed liquid damper concept
Lastly, a damping component which is unavoidable and must be incorporated, is the internal
motor friction. As stated previously, this is a non-ideal, non-linear damping force that is apparent
when a motor is accelerated from a stop, requiring above a threshold level of current before the shaft
begins to spin. This is a source of damping, however, and will be quantified with motor
11
characterization. Efforts will be made to ensure this does not greatly deter from ideal system
performance.
Mass Alternatives
Prior Designs of the take-home kit have incorporated a 1/2” bore off-the-shelf shaft collar as
the primary rotational system mass. While cheap and effective, this mass relies on a magnet to be held
to the rotating pulley, and allows for no flexibility to view the effect of mass changes on system
response. In its place, another alternative presently under consideration is the use of stacked
washers. These are a very cheap alternative that would allow for many levels of mass configuration.
However, a new method must be designed to secure them and prevent movement and sliding, which
will create inconsistencies in system performance.
Position/Velocity Sensor Alternatives
One of the most important parts of this kit is the ability to detect rotational position. Along
with this requirement, a new requirement has been introduced for the 2013 take-home team to also
add velocity awareness to the kit. Velocity is the first derivative of position, so any sensor which
measures position can be used to derive velocity in software.
The 2006 take-home kit revision incorporated a potentiometer to detect position. To do this,
one potentiometer terminal is provided a voltage signal, one is provided ground, and the wiper is the
output signal. This is used in a manner of voltage divider circuit, with output voltage varying with
position. This solution is reliable and performs well, however, there are a few disadvantages. First,
it requires a geared, belt, roller, or other type of indirect connection to the motor, as it is not easily
packaged into the motor’s rotational axis. That connection is sensitive and problematic. Second, to
obtain a quality, low-friction potentiometer that is not limited by number of turns, it is quite expensive
(around $15 each). Lastly, the potentiometer is large and with moving parts, does have the potential
to eventually fail. See Figure 1-9 for the 2006 configuration of the potentiometer.
Figure 1-9: Prior potentiomer configuration, with gear pulley
One proposed concept to provide position measurement is through the use of a
phototransistor. An example of this type of component is shown in Figure 1-10. This device emits
infrared light, and measures reflections of the signal. It could be used in conjunction with a wheel
that incorporates evenly spaced black/white color transition, or with a wheel which has evenly spaced
holes around the circumference directly above the mounting point of the sensor. Doing this will
provide a measurement best used for velocity, but can be integrated to find position. This solution is
12
simple and cheap. However, it comes with a few drawbacks; primarily the lack of built-in position
measurement, which will create a more complicated programming sequence and larger error. The
second major drawback is resolution which is limited by the total number of black/white transitions or
holes.
Figure 1-10: Phototransistor (courtesy of sparkfun.com)
A final position/velocity concept under consideration is that of a Hall Effect type of system.
This system uses a magnet, which is mounted to and rotates on the central shaft, and two analog Hall
Effect sensors, similar in size and cost to the phototransistor, which are mounted some distance away
and located with 90 degrees of rotational separation. See Figure 1-11 for a schematic of this
configuration. An alternative could be to mount the magnet offset some distance from the axis of
rotation, which triggers a Hall Effect sensor as it rotates. This is undesirable, however, due to the
negative impacts on rotational inertia of the unit and the creation of vibration during operation.
The Hall Effect sensors make a measurement on the strength of the magnetic field created by
the magnet. As it rotates, the magnetic field’s magnitude changes and polarity reverses, which is
captured by the sensors. The use of two sensors provides precise position information about the
rotational position of the system, as well as measurement sensitivity throughout the range of rotation.
This is a standard, well known way of making position measurement, and has been tested to perform
well.
This concept is very promising as it is very cheap, with the sensors themselves priced at
around $1 each. The magnet being relatively cheap as well. It is reliable and robust, as the sensors
incorporate no moving parts and are very accurate. All that is required is solid mounting of the
magnet to the rotational shaft, and a precise mounting points and orientation of the sensors.
Figure 1-11: Hall effect sensor configuration, with north and south magnet poles in blue and red,
respectively
13
Software Package Alternatives
The selection and creation of a software package, by which the student will perform the lab
tasks and view system output, is a vital part of the success of this take-home lab kit. In the past, the
desktop software that the student uses to interact with the physical system was implemented in Visual
Basic and pre-compiled before being run by the student. This created many compatibility problems
with various operating systems and software environments. To rectify this, the 2009 iteration of the
take-home kit made use of a graphical user interface implemented in MATLAB. As part of the initial
requirements for the 2013 revision, it was requested that the software be implemented in a platform-
independent software package, which makes installation and use on student home computers less
problematic. In addition, it is requested that the software be easily updated and enhanced, by TA’s or
professors, through the use of straight- forward implementation and thorough documentation. Lastly,
the use of the Arduino as the hardware control device limits the ‘server’ software to only being
Arduino-compatible C-code.
The first software package concept is using a web-based user interface to interact with the
serial port to which the Arduino is connected. There is no straightforward way to do serial
communication via HTML or JavaScript, so a local web server would have to be developed which
interacts with a serial driver to control output with the Arduino. This solution is convoluted and
complicated, and probably isn’t easily developed in the time allotted for the semester.
The second software solution consists of interaction with the Arduino through serial
communication in MATLAB Simulink. This would allow the software to be implemented in the
function-block style Simulink development environment, which generates the MATLAB side, and
Arduino server-side code to execute the Simulink model. This is a reasonable alternative and allows
easy enhancements and changes to the software and user interface. However, professor Durfee
expressed concern with adding another level of abstraction onto the software package by introducing
Simulink. However, using MATLAB as the user interface software is very viable; students may
obtain an educational license for MATLAB for free from the university, and this software is already
used in ME3281 for homework assignments. Using it for the take-home lab exercises would provide
seamless integration with current curriculum. MATLAB is also platform independent and would be
less sensitive to different operating systems and system environment differences.
Similar to the previous option, the final option is to write a custom MATLAB user interface
and/or modify the 2009 MATLAB software to fulfill requirements for the 2013 implementation. This
combines all the benefits of using MATLAB with the simplicity of eliminating Simulink from the
equation. This software will be well documented and laid out for easy updates and changes in the
future. It will be combined with custom Arduino server code that will control the hardware and
provide position and velocity feedback to the user interface for graphing and analysis. It will be
possible to export data to a CSV file for manipulation in other software packages. However, one
downfall is the reduced performance of the serial communication with the Arduino due to the
overhead of the MATLAB environment. However, with careful benchmarking and testing, the impact
of this could be minimized to produce a useable, flexible user interface for the take-home kit.
14
1.5 Concept Selection
System Type
Selection Criteria Rotational Linear
Cost 0 0
Complexity 0 0
Size 0 -
Custom Components 0 -
Component Selection
Motor 0 -
Spring 0 +
Mass 0 +
Damper 0 +
Position 0 -
Net Score 0 -1
Table 1-3: System Type Selection
In Table 1-3, the rotational system, because it was used in the past, is set as the benchmark. The cost
criterion refers to the price to purchase the kit as a whole, including consideration to all components.
Because cheaper is better, points are awarded for a lower cost. Complexity, size, and number of
components are to be minimized, and points awarded to this end.
Using a Linear type of system eliminates some of the complexities and challenges of
rotational; specifically with regard to spring and damper implementation. However, the size and
number of custom components required to allow smooth motion of a linear system is also a
formidable challenge. One of the custom components required is a necessary type of cam drive in
order to turn the rotational motion of the motor into linear motion. In the end, a linear system also
makes a PID control lab, as well as velocity control, quite difficult. Both alternatives have strong
points, but in the end, staying with a rotational system makes the most sense.
15
Spring
Selection Criteria Rubber
Band
Torsion
Spring
Split
Mount
Coil
Spring
Shared
Mount Coil
Spring
Cost 0 - - -
Size 0 + + -
Linearity 0 0 - +
Range of Motion 0 - - 0
Working Life 0 + + +
Mounting 0 - + +
Net Score 0 -1 0 1
Table 1-4: Spring Selection
The criteria used in Table 1-5 include a few that are to be minimized; cost and size. To be
maximized are working life, linearity, and range of motion. The mounting category awards points to
the solutions that are easiest to mount.
Using the rubber band spring configuration as the baseline, the other spring options offer
some attractive benefits. The torsion spring, however, must eliminated due to the inability to rotate in
both directions. It does seem physically possible, but the manufacturer recommends against it, and
likely would lead to non-linear spring force. The shared mount coil spring offers all the benefits of the
split mount coil spring, but also improves by allowing a larger range of motion, and better linearity.
While it suffers from size, this configuration is determined to be the best moving forward.
Damper
Selection Criteria Internal
Friction
Rotary
Damper
Liquid
Fan
Damper
Cost 0 - 0
Size 0 0 -
Linearity 0 + +
Cycle Speed 0 - +
Damping
Coefficient 0 + +
Net Score 0 0 2
Table 1-5: Damper Selection
The damper selection criteria includes cost and size which award points for a decrease from
the internal friction benchmark. Linearity, cycle speed, and damping coefficient are increased when
points are awarded.
It is concluded that internal friction is unacceptable as the only source of system damping,
due to its non-linearity. There are various forms of commercially available rotary dampers, but in the
end this option must be eliminated due to the restriction of 10 rotation cycles per minute; it is desired
16
to have the take-home kit operate at higher frequencies than that. This leaves only the liquid fan
damper, which offers a great opportunity for hands-on education about damping for ME3281 students.
This type of damping is not quite ideal, but will be sufficient and helpful for the purposes of this kit.
Position/Velocity Sensor
Selection Criteria Potentiometer Phototransistor Hall Effect
Cost 0 + +
Size 0 - 0
Accuracy 0 - +
Position 0 - +
Velocity 0 + 0
Simplicity 0 + -
Net Score 0 0 2
Table 1-6: Position/Velocity Sensor Selection
The criteria used in Table 1-7 include accuracy, velocity, and simplicity, which award points
to an increase from the benchmark potentiometer. Size and cost are to be minimized. The position and
velocity criterion refer to the capability for accurate measurement of those metrics by the design under
consideration.
The potentiometer that has been used in past revisions of the take-home kit has been a good
solution to position sensing. However, there is room for improvement with regard to cost and size.
Use of a phototransistor and trigger wheel offer a cheap, simple alternative, however, this is mostly
only good for velocity measurement and suffers from a lack of resolution. The 90 degree separated
Hall Effect sensors with shaft mounted magnet offer a great overall solution; good accuracy and
resolution, precise rotational position measurement. Even though this will require more programming
and a computational derivative to be calculated to find velocity, it is the solution that will be
implemented in the 2013 take-home lab kit.
Software Package
Selection Criteria MATLAB Simulink Web-Based
Cost 0 0 0
Simplicity 0 - -
Ease of Use 0 - 0
Ease of Maintenance 0 0 -
Platform Independence 0 0 -
Ease of Troubleshooting 0 0 -
Relation to Course 0 0 -
Net Score 0 -2 -4
Table 1-7: Software Selection
The criteria used to evaluate the software are mostly self-explanatory. Ease of maintenance
refers to the difficulty involved with fixing bugs and making changes, which is also related to ease of
troubleshooting. Relation to course awards points for using packages that are used elsewhere in
17
coursework, and platform independence awards points to solutions that can be run on a larger number
of operating systems.
The 2009 version of the take-home kit implemented a MATLAB User interface, used with a
USB driver to communicate with the PIC board. This approach is used as a baseline for 2013
selection. Simulink offers a good off-the shelf method for developing a control program with both
Arduino server-side and a user interface. However, this introduces another layer of abstraction to the
software, and if configuration is necessary for the Arduino code that is generated, it could prove
difficult. A web-based approach would be very easy to use and could be run on a host computer
without any special software (other than a web browser), however there would be many other issues
such as the necessity of a web server or other middleware to interact with the serial port and the
difficulty with troubleshooting or modifying such a setup. Lastly, this solution would not be exempt to
compatibility issues due to the wide range of browsers and versions that might be installed on a host
computer, would exponentially complicate implementation, and ultimately require an ‘approved
browser list’ which would have to be tested and updated over time.
In the end, it makes the most sense to stay with a MATLAB user interface, as it is the most
platform independent, would be easy to modify and troubleshoot, is free to use and very familiar to
students, and ties in well with other MATLAB assignments in ME3281. The features implemented by
the 2009 group can be leveraged and extended to provide a full feature set for the take-home kit.
2. Design Description Supporting Documents
2.1 Manufacturing Plan
2.1.1 Manufacturing Overview
Great consideration was taken in the design of this lab kit to meet two main criteria: low cost
and simplicity. To keep the costs low, only two pieces of the kit are custom made, the shaft coupler
and the PCB. The rest of the kit utilizes off-the-shelf parts that are mass produced, widely available,
and low cost.
The pulley assembly is the centerpiece of the whole kit, and is responsible for providing a
means for mounting the mass, springs and magnet on to the motor. This assembly consists of a shaft
coupler, an off the shelf pulley and a magnet. The shaft coupler is a one inch long, plastic cylinder
with a 5/64 inch hole drilled in one end, to a depth of 3/8 inch. The magnet is then attached to the
coupler, at the end with the hole, with glue, and the pulley is attached likewise to the coupler, just
above the magnet. When the coupler is glued to the motor shaft, the magnet is in the optimal position
for use with the Hall Effect sensors. In addition, the pulley is in position to be used as the mounting
point for the springs.
The springs are mounted by wrapping a string around the pulley. The string is cut to a
precise length and has one end tied to each spring. The springs are then attached to two mounting
points on the PCB. The string is held to the pulley by first double wrapping it around the pulley, and
then securing it with a tight rubber band. Above the pulley the remainder of the shaft is used to mount
the washers. The washers are secured by a tight press fit connection and also by a rubber band
wrapped on the shaft above them. This connection ensures that the washers do not slip during
operation of the kit.
The design of this assembly is created with the principle of cost reduction, in mind. The
vertical stacking orientation of all of the components minimizes the amount of space that the kit
occupies and therefore reduces the size of the PCB to which the components are mounted, and the
18
size of the box needed to contain the kit in its entirety. All of this supports the goal of reducing the
cost of the kit, and maintaining a standard of simplicity.
The custom PCB contains only a few holes to mount the motor, legs, and springs, along with
traces for the Hall Effect sensors’ power, ground, and signal terminals. Precise spacing is used in the
Hall Effect sensor holes, to maintain a correct distance to the magnet. These are shown in Figure 2-2,
as the 3-hole groupings above and to the left of the motor through-hole and attachment holes. The
spring attachment points are provided as holes on the bottom of the figure, and attachment holes are
provided at each corner, for the legs of the kit. This is a simple design, requiring only a 1-layer PCB,
though more layers may be used. The configuration also acts as the lab kit base, further reinforcing
the low cost and simple design philosophy. The PCB has been designed using the ExpressPCB
software from www.expresspcb.com which allows for easy and intuitive PCB creation, instant quotes,
and fast order turnaround.
2.1.2 Part Drawing
The main component, the pulley, is displayed below in Figure 2-1.
Figure 2-1: Line Drawing of the Mass System
The PCB schematic is displayed below in Figure 2-2.