Critical Design Review


1
Critical Design Review
Team RIDE

• Brennan Dayberry
• Adam Marrapode
• James McGlynn
• Ben Sufit
• Chris Taylor

2
Presentation Overview

• Project Visual Walkthrough
• Problems and Design Revisions
• Revised Mechanical Functions
• Flow of Information
• PC Plug-in
• FPGA Game Code and Motor Control
• FPGA NiosII and Controller Board
• Motors Drivers and Power
• Schedule
• Division of Labor
• Budget Considerations

3
Visual Walkthrough
4
Problems and Design Revisions

• Mechanical Engineering
• We completely underestimated the amount of
  mechanical engineering knowledge required to
  implement our project
• Mechanical engineers have consulted with us on
  various topics relevant to our project and
  helped guide us in the right direction
• Due to the machining metalwork and welding
  involved a few of us have gained access to the
  ITLL machine shop and have a mechanical engineer
  with welding experience lined up to help us with
  any welding needed
• We have attempted to make our mechanical design
  as simple as possible while maintaining the
  overall goal of our project

5
Problems and Design Revisions

• Linear Actuators
• Grant Immohara from Mythbusters provided
  information on actuator suppliers
• Very hard to find exact actuators to meet design
  specifications for load forces and actuator
  speeds
• Way too expensive cannot accommodate within our
  budget
• Much harder to implement into design than
  originally perceived
• Overall Linear actuators were a bad idea and we
  wasted too much valuable time trying to find the
  perfect one

6
Problems and Design Revisions

• Motors
• After researching variable AC DC and stepper
  motors we decided to use stepper motors and
  salvaged two from the previous capstone project
  the BP gas line robot
• Well within our budget as they were free and
  higher models are within budget (300-400)
• Stepper motors much easier to implement into
  design both mechanically and electrically
• Feedback easier via step control and shaft
  encoders
• Use crank arm and levers to achieve 2 degrees of
  freedom
• Stepper motor system already has AC-DC converter
  power supply

7
Problems and Design Revisions

• Overview
• Tremendous amount of mechanical input was needed
  from outside sources
• Calculations of Torques and Force loads proved
  that linear actuators are not efficient but
  proved that stepper motors could carry out design
  specs by using a crank and lever system
• Metal support frame for cockpit was deemed to
  heavy so we went with plywood instead and will
  improve rigidity with perimeter molding of
  angle-iron
• Motion base will have 2-degrees of freedom with 2
  stepper motors both on the front edge.

8
Mechanical Functions

• When motors move in unison together cockpit
  pitches up and down
• When motors move opposite directions cockpit
  rolls side to side
• Approximately 7-12 range of motion in both
  degrees of freedom

9
Universal Joint

• Salvaged from an old truck drive shaft
• Placed at center of mass to support majority of
  weight
• Restricts motion to two degrees of freedom
• Adjustable height

10
Ball Joint

• Allows levers to move freely at pivot point on
  cockpit
• Prevents lever arms from bending when the cockpit
  moves
• Easily available from local suppliers

11
Torque Requirements

• Need 15 pounds-force along front edge per motor
  (empirically measured)
• Crank arm is 2 inches long
• We need 30 inch-pounds or 3.4 Newton-meters at
  motor shaft

12
Cockpit Pitching

• Both motor crank arms move together so cockpit
  pitches up
• The delta distance between the cockpit old
  position and new is about 2.5 inches
• Allows pitch swing of about 5 inch delta or 6

13
Cockpit Rolling

• Motor crank arms move in opposite direction so
  cockpit rolls to the left
• Delta distance between old and new position is 2
  inches
• Provides roll swing of about 4 inch delta or 12

14
Information Flow

• PC Plug-in filters game data into movements then
  sends to FPGA Game Code Module
• Game Code Module considers current positions
  evaluates new desired positions based on
  feedback then sends data to FPGA Motor Code
  Module
• Motor Code Module converts simple movement
  commands into pulses and directions for each
  motor
• Motor Drivers take PWM signals and drive motors
  correspondingly

15
PC Plug-In Overview

• Dynamically-Linked Library
• Routines run in real-time when race is started
• Telemetry data held in game data structures (C
  classes)
• Telemetry updates at 91 FPS
• Telemetry converted to motion profiles
• Structures - Filter - Serial Port

16
PC Plug-In Raw Data

• Data polled on frame update and sent to filter
  functions
• Data sent in three types
• Acceleration forces created by car movement
• Impact forces created by impact with
  environmental hazards (walls other cars)
• Pitch/Roll forces created by movement of car in
  relation to normal (includes bumps jumps)
• Data in Cartesian coordinates

17
PC Plug-In Filters

• Filter for each type of data
• Data sampled and converted at 5 Hertz
• rate variables held as constants to allow for
  easy optimization with hardware
• Each filter converts data to simple motion
  profile
• Simple motion profiles sent to superposition
  function
• Superposition function sends final profile to
  serial port

18
PC Plug-In System Functions

• initialize
• Initializes communication with controller board
• log
• logs any pertinent data (meant for development
  debugging)
• send_profile
• sends direction and magnitude of motor arm motion

19
FPGA Game Decode

• Inputs from PC
• Feedback via variables and optical encoders
• Output to Motor Control
• cmdmov values (command move)
• Magnitude and Feedback
• Sample Code
• Early feedback
• Output

20
(No Transcript)
21
Inputs Feedback and Outputs

• The input to the FPGA will be a set of two
  integers corresponding to the desired movement
  and magnitude.
• The feedback is the current position of the
  motor provided by shaft encoders and known
  variables.
• The output gives the command to the motor based
  on the input and feedback

22
(command move) cmdmov values

• 1 front (brake)
• 2 back (acceleration)
• 3 front right
• 4 front left
• 5 back right
• 6 back left
• 7 CRASH or rumble strip (depending on magnitude)

23
Magnitude and Feedback

• Magnitude is a value of one to ten determined by
  the PC plug-in.
• Feedback is a number between zero and two hundred
  (200 steps per revolution) that contains the
  current position information.
• The code calculates which movements to make and
  determines if the movement is possible based on
  the feedback

24
Sample Code Nested Switchor Lookup Table

• switch(cmdmove)
• case 1 //pitch front
• swich(mag)
• case 1 // magnitude of one if ((fdbk -
  mag)
• then mag fdbk
• movemoter1(cmdmov mag)
• movemoter2(cmdmov mag)
• break//moves motor as far as pos
• case 2 // magnitude of two
• if ((fdbk
  mag) 200)//check

25
FPGA Game Output

• The FPGA Game Code module sends basic movement
  information to the FPGA Motor Control module to
  be converted into real movement
• The Game Code module handles all feedback
  decisions so that the Motor Control module only
  has to translate basic movements into motor
  rotation

26
FPGA Motor Control

• Translates fundamental movements into number of
  steps and direction of rotation of both motor
  shafts.
• Only provided with the changes in movement from
  the last position to make translation fast and
  simple.
• Relates desired cockpit behavior into necessary
  motor behavior through basic physics
• Controls logic driving the PWM and direction
  circuits on FPGA

27
Altera Nios II Softcore Design Flow
3
2
SOPC Builder (System-on-a-Programmable-Chip)
Nios II IDE (C/C)
Quartus II (VHDL/Verilog)
4
1
5
FPGA
28
PC
RS232
Position Feedback
Logic
UART
Cyclone FPGA
Motor Driver
29
Control Board Overview
30
Breakout Board from www.devboards.de

• Connects FPGA to PCB with headers
• Simplifies FPGA mating to PCB
• Provides JTAG and Serial Programming Interfaces
• Allows for easy replacement of FPGA

31
Some Schematics
32
Some Schematics
33
Slo-Syn Stepper Motor

• 5.44 N-m Minimum Holding Torque
• 2 phase
• 1.8 per step (200 steps)
• Controlled via MD808 motor driver
• We will be operating at 60-100 rpm at 5 N-m
  continuous torque

34
MD808 Motor Driver

• Opto-isolated control
• 80VDC 4A power
• Control via pulse signal one rising edge is one
  motor step
• Direction control high-low signals CCW or CW
  rotation direction

35
Feedback

• Because we made the design choice to go with
  stepper motors instead of liner actuators we can
  get some early feedback by keeping track of the
  approximate position of the motors.
• We anticipate that this will not be sufficient
  for the finished product and plan on using
  optical shaft encoders to provide the exact
  position of the motor.

36
Motor Power Circuit

• Converts 120VAC to 80VDC
• Transformer converts 120Vrms to 60Vrms
• Full wave rectifier supplies 80Vpeak full wave
• Filter capacitor turns full wave into 80VDC output

37
Revised Schedule
38
Deliverables
39
Division of Labor
Ben
Adam
James
Chris
Brennan

X

X
X
X
Mechanical Design
X
X
X
X
X
Mechanical Assembly
X
X
X
X
X
Motor Control
X
X
X
X
X
Motor Feedback

X
X
X
Programming PC Driver
X
X

X
X
FPGA Game Code
X
X

X
X
FPGA Logic
X
X
X

X
FPGA NiosII Softcore

X
X
X
X
Communications
X
X
X
X

Analog Circuits
X
X

X

PCB Layout
X

X

X
Power
X
X
Budget

40
Budget Overview

• New Investors
• Joseph Brennan owner of Southern Lending LLC
  Investment Firm
• Invested 1000.00 towards project through CU
  Foundation
• www.southernlendingllc.com
• DENT Sport Garage
• National Pro Rally team and automotive upgrade
  shop
• Will support safety of project
• 7-point safety belt harness system neck collar
• Better and newer seat for larger individuals
• www.dentsport.com

41
Budget Overview

• Overview of Materials and Parts
• NEOS II FPGA
• Devboard (free!)
• Breakout FPGA (180)
• Two DC Stepper motors and drivers
• (free!)
• If unable to repair defective driver new one
  will cost 365
• AC-DC Converter Power Supply
• (free!)
• Miscellaneous scrap materials purchased thus far
• U-bolts plywood driveshaft for center pivot
• 150.00
• Total Spent 330.00
• Well within budget requirements for now

42
Questions