Transcript CLOSING THE LOOP - University of Southern Maine

CLOSING THE LOOP
Teaching Control System
Theory and Design Through
Experimentation
Project Design Engineer
Project Mentors
Mark Kapsch
Dr. Brian Hodgkin
Mr. Steve Innes
Preliminaries
Purpose of Project
 Overview of Control System
Theory
 Antenna Azimuth Model
Construction
 Parts and Labor
 Scalability

Reason for Design

Experimental Model
for Control Systems
Design Course,
ELE425
• Commonly
Encountered Control
System Design
Problem
• Experimentation and
Analysis Enhances
the Learning
Experience
70m Deep Space Tracking Antenna
GoldStone Complex, California
Photo Courtesy of NASA JPL
Basic Control System Theory

Control Systems Use Feedback
Loops to Manipulate the Circuit
Output
Desired
 s
Azimuth i  
Angle
Input
Potentiometer
K pot
Vi  s 


Ve  s 
Differential
Amplifier
K
Motor and Load
Ea  s 
Km
s  s  am 
Position
Sensor
K pot
m  s 
Gears
Kg
o  s 
Antenna
Azimuth
Angle
Transfer Function for the Model
Antenna Azimuth System
T s
 s
i
KK pot K m K g
s  am s  KK g K m K pot
2
o  s 
Building the Experimental Model
As Constructed
_
+
Model Antenna
Azimuth System
assembled and
ready for testing
Analog System Schematic Diagram
Antenna Azimuth Control System
University of Southern Maine
Mark Kapsch
04 May 2007
VSS
12 V
Rcomp
50%
Vin
50%
RG_10
60%
7
RG_90
4.7kOhm_5%
1
5
U1
S1
3
RF2
10kOhm_5%
6
M
2
LM741CN
MOTOR
4
Vfb
50%
RG
RF
5.1kOhm_5%
10kOhm_5%
Gain = RF/RG
VEE
12 V
Proposed
T
2006
2007
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
Produce Classroom Materials
Building
Programming
I
Pre-Proposal
Due
Parts on Order
Algorithm
Complete
Unit Built
Test and
Adjustment
Project
Proposal Due
Control System Design Analysis and Specification
M
Presentation
E
L
I
N
E
Actual
2006
2007
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
Building
Second
Pre-Proposal
Complete
Project Not
Unique
First Pre-Proposal
Complete
Design Analog
Control System
All Parts
On-Hand
Power
Supply
HandyBoard Motor Inadequate
Preamp
Outputs Are Stepped
Not
Motor
Working
Torque
Control System Design Analysis and Specification
Project
Proposal
Complete
10 56

56 56
Test and
Adjustment
Presentation
Parts and Labor

Initial Budgetary
Allocation
• $1000 or less

Part Costs

Labor Costs
• 217.5 Hours X
$16.00/Hour =
$3480
• Estimated

$745.60
• Actual

$909.64
• Overrun

$164.04
• Total $4353.86
Scalability of Design
Built-in Option for Digital Control
System Implementation
 Ability to add Sensors or Motors
above the Rotating Pedestal
 Modular Design

• Allows alternate components to be
easily implemented
Antenna System
Phase 2
Project has been selected by
Gabe Garza for enhancement
 Elevation system to be added
 Modification scheduled for Fall
07 to Spring 08

Summary
Need for Project
 Control System Theory
 Development and Construction
 Project Cost
 Future Growth

Questions?
Backup
Block Diagram Reduction
Motor and Load
Potentiometer
 s
K pot
i
Vi  s 

Ve  s 
Power Amp
Ea  s 
K

Gears
 s
Km
Kg
s  s  am 
m
o  s 
Potentiometer
K pot
 s 
Potentiometer
i
K pot
Vi  s  
Ve  s 

KK m K g
o  s 
s  s  am 
Potentiometer
T s 
K pot
 s

i
KK m K g
K pot
s  s  am 

o  s 
G s
 s

i

K pot KK m K g
s  s  am 
G  s
1 G s H s
T s
o  s 
 s
i
KK pot K m K g
s 2  am s  KK g K m K pot
o  s 
Unknowns
Motor
J a , Da
N1
N2
JL
DL
Mechanical System
m  s 
Ea  s 

Kt
Ra J m
Km

s  s  am 

Kt Kb  
1 
s s 
 Dm 

J
R
m 
a 

Kt
is the motors torque constant divided by the motor armature resistance and
Ra
is equal to the rated stall torque of the motor divided by the motor armature input voltage
TStall
.
ea
The value
Armature input voltage divided by the no-load speed of the motor results in the back
ea
EMF constant of the motor or Kb 
.
noload
The Mechanical System Problem
 N1 
Jm  Ja  J L 

N
 2
2
 N1 
Dm  Da  DL 

N
 2
2
The Electrical System Problem
K pot
24volts

 3.82
2 rad
RF
K
RG
This calculation makes two assumptions. First that saturation is never reached and second
that the response of the amplifier is fast enough to neglect delay effects. If it is
determined through experimentation that the response has been delayed the transfer
K
RF
Rf
function for the amplifier will need to be modified to
where K 
and a 
.
sa
RG
RG