Contract EHSH 71-002
S/N 17220.00
ANALYSIS AND ADVANCED
DESIGN STUDY
OF AN
ELECTROMECHANICAL
TRANSMISSION
APRIL 1971
Prepared for
OFFICE OF AIR PROGRAMS
ENVIRONMENTAL PROTECTION1 AGENCY
Ann Arbor, Michigan
TRW
srirfM] cuouf
am VACC PARK • PCOOHOOHCACH CALIFOHNIA son*
-------
Contract EHSH 71-002
S/N 17220.00
ANALYSIS AND ADVANCED
DESIGN STUDY
OF AN
ELECTROMECHANICAL
TRANSMISSION
APRIL 1971
Prepared for
OFFICE OF AIR PROGRAMS
ENVIRONMENTAL PROTECTION'AGENCY
Ann Arbor, Michigan
TRW
SYSTf/HS GROUP
ONf SPACC PARK • FICOONOO BEACH. CALIFORNIA
-------
ANALYSIS AND ADVANCED DESIGN STUDY
OF AN
ELECTROMECHANICAL TRANSMISSION
April 1971
Prepared for the
Office of Air Programs
Environmental Protection Agency
Under Contract No. EHSH 71-002
by
TRW Systems Group
One Space Park
Redondo Beach, California 90278
S/N 17220.000
-------
FOREWORD
This report contains a summary of work performed by TRW Systems
for the Office of A1r Programs, Environmental Protection Agency, under
contract EHSH 71-002.
Dr. George H. Gelb was responsible for overall project direction
and report editing. Mr. Baruch Berman was responsible for the design
and development of the electrical systems. Dr. Eli as Koutsoukos was
instrumental in developing the emission testing procedures and instru-
mentation. Dr. T.C. Wang provided the computer analysis of the elec-
trical system and played a role in overall system analysis. Other
members of the technical staff who contributed to the work include:
Messrs. H. Gehm and L. Inouye - electrical subsystem, Mr. W. Potter -
engine control and instrumentation and Messrs. E. Hoover and R. Lewis -
technical support.
-------
TABLE OF CONTENTS
Summary and Recommendations 1
Introduction 6
1.0 The Electromechanical Transmission (EMT) 8
Description of Concept and Modes of Operation 8
System Functional Requirements . 13
2.0 Torquer System 14
Review of Candidate Torquer Systems 14
Selection of Torquer System 20
Time Ratio Control (TRC) Systems 20
Selection of Power Pass Elements for TRC 28
Trade-off and Selection of a Power Control Unit (Torquer PCU). . 36
Description of EMT Torquer System 60
3.0 Speeder System 65
System Requirements and Selection Criteria 65
Selection of the Speeder 65
Rational of Speeder Operating Mode and PCU Selection 65
Speeder PCU 67
4.0 Rotating Machines 70
Torquer Selection 70
Torquer Specifications 70
Speeder Specifications 71
5.0 Dynamometer Testing of the EMT System 73
EMT Dynamometer Equipment 73
Instrumentation and Data Processing 73
Computation Subroutines 81
Computer Print-Out and Data Presentation 94
Data Logging 95
EMT Dynamometer Performance on LA-4 Route 95
Statistical Representation of EMT-Dynamometer Operation .... 104
Summary of EMT-Dynamometer Testing 120
6.0 Advanced Design Study of Torquer System 124
Torquer Drive System Analysis 124
Torquer Design 138
Torquer Design Improvements 141
Speed Charging of the Torquer 145
Chopper Frequency Selection and Power Control Improvements ... 146
Torquer System Failure Analysis 149
7.0 Advanced Design of the Speeder System 161
Review of Speeder Performance on LA-4 Tests 161
Machine Losses 161
Speeder PCU Improvements 170
Preliminary Design of an Advanced Speeder PCU Control Circuit. . 179
Speeder System Failure Analysis and Protection 182
-------
Page
8.0 Advanced EMT Power Train Operational Concepts 190
Solid State Switching System . . 190
EMT Power Train Start-Up Concepts 202
Override Options to the EMT System 204
9.0 Effect of Selected Parameters on Exhaust Emissions from an
Internal Combustion Engine 206
Introduction 206
Determination of Engine Emissions During Cold and Hot Operation 206
Effectiveness of a Three-Component Catalyst on Exhaust Emission 228
Prediction of Emission Signature of an EMT System with a Three-
Component Catalytic Drive 228
References 231
A-l Comparison of Boost-Buck and Boost PCUs for Regeneration . . .232
A-2 Mathematical Model of Torquer System 236
-------
NOMENCLATURE
The following set of symbols and definitions are used in the report.
In places where other symbols are used, they will be defined at that
point and are to be used only for the section in which they appear.
V - instantaneous value of a voltage, volts
I - instantaneous value of a current, amperes
t - time (units either omitted or presented under particular usage)
T - period of chopper operation, seconds
f - frequency, 1/T, hz
L - inductance, henrys
C - capacitance, farads
R - resistance, ohms
E - generated voltage of a machine, volts
S - denotes a saturable reactor
D - denotes a diode
SW - denotes a switch
F - denotes a fuse
F,R - denotes forward or reverse direction
Subscripts
c - chopper; commutation
g - generator
T - Torquer
B - battery
S - Speeder
on - power pass element conducting
off - power pass element not conducting
Greek Letters
a - ratio of power pass element on time to total chopper period;
WT
6 - ratio of commutating SCR time on to total chopper period
Averaging: a bar (-) over any quantity g means
= f g(t) dt
J
_
defined time interval
i i i
-------
Summary and Recommendations
TRW Systems Group has performed a study entitled "Analysis and
Advanced Design Study of an Electromechanical Transmission" under con-
tract EHSH-71-002 for the Air Pollution Control Office of the Environ-
mental Protection Agency. This report presents the results of work done
during the period of the contract from July 13, 1970 to April 26, 1971.
The work performed falls into two major divisions. The first portion
is devoted toward eight tasks. These tasks were to:
• Evaluate the performance data of the Electromechanical Transmission
(EMT) operating on a dynamometer
t Perform engineering trade-off studies on the power conditioning
units of the Speeder (generator) and Torquer (traction motor)
subsystems and to present up-graded design criteria
• Assess the performance of the Speeder and Torquer machines and
to present design criteria for advanced versions of both machines
• Develop advanced operational concepts such as start-up techniques
and means for up-grading the transfer of the Torquer system between
the motor and generator operating states
• Develop analytic models of Torquer system operation and to use
these models to assess the effect of design parameters.
The second portion of the study was intended to provide operational
data on the performance of an internal combustion engine (ICE) when used
within the context of a hybrid power train. Of particular significance
is the determination of the role of the cold period of operation to total
engine emissions. Preliminary data on the effectiveness of a catalyst on
engine emissions was gathered.
All of the experimental work was gathered from a breadboard version
of the EMT. The equipment was built to prove the EMT principle, and has
performed satisfactorily as a basis for assessing the performance of a
hybrid system designed for low engine emission operation. The units
which comprised the breadboard EMT were assembled rapidly on the basis of
their availability. None of the rotating machines or gearing was optimized
to any degree. Therefore, it must be recognized that the breadboard EMT's
performance is indicative of first effort equipment and major system
improvements are possible.
1
-------
The EMT was operated on the dynamometer and driven over the LA-4
route, a route typical of driving in the Los Angeles area (Section 'J.l).
Various operators, using strip chart driver's aids, drove the equipment
under dynamometer settings representative of vehicles weighing 2000 to
4000 pounds. The dynamometer testing revealed the following:
• The overall transmission efficiency of the EMT for LA-4 operation
is approximately 50%. Efficiency as defined as the demanded road
load energy divided by the engine shaft output energy (Sections
5.6 through 5.8). Fuel consumption is about 18 miles per gallon.
• Major inefficiencies can be associated with each of the major
subsystems which were purchased or built v/ithout time for detailed
design trade-offs--
- the Torquer subsystem is inefficient due to the impact of
the chopper on the current flow through the Torquer
- the Speeder is a lossy machine whose efficiency is further
degraded by its operation into a rectified and chopped load
- the gear box is a primary source of inefficiency—the bread-
board gear box is greatly overdesigned, with excessive
lubrication and unnecessary gear losses.
- the component efficiencies should be improved greatly by re-
design and selection of equipment for the specific type of
EMT service.
t Battery impedance impacts heavily on the overall utilization of
engine power. It limits the maximum charge and discharge powers
and accounts for a major-source of energy loss.
0 The battery receives almost equal charging energy from the Speeder
and Torquer and thus the generator efficiency of the latter must
be improved.
e The effectiveness of regenerative braking in providing battery
charge power is small with respect to the contribution coming from
the engine. Means should be evolved for increasing its effect by
improved Torquer generator operation and redesign of the braking
controls and control limits (Section 5.8).
• All major subsystems with the exception of the Speeder functioned
in ambient air without forced cooling during tests on the LA-4 cycle.
Computer studies of the Torquer drive system have verified the import-
ance of the selection of chopper frequency and machine inductance on Torquer
system efficiency. The filtering effect of the armature circuit inductance
impacts the system performance in two ways. With increased inductance the
-------
resistive losses in all elements of the circuit decrease and magnetic
losses in the machine are reduced. , Increased chopper frequency decreases
the ohmic losses of the system (Section 6.1).
Two methods are available for increasing the Torquer efficiency; the
addition of armature circuit inductance by increased series windings or an
external choke, and the lamination of the flux carrying elements of the
machine. Experimental evidence shows that both approaches increase the
machine efficiency significantly (Section 6.3). Further economic and
engineering trade-offs are required to determine the best approach or
combination of approaches.
Losses in the Speeder system can be reduced by improved aerodynamic
design and a lower operating speed, the latter with some weight penalty.
Major hysteresis and eddy current losses are introduced by operating the
alternator into a rectifier and chopper. The resulting current waveforms
contain 3rd, 5th and 7th harmonics which can be greater than 10%. An
advanced Speeder design should incorporate a higher flux gap density and
low electrical conductivity pole faces (Section 7.2).
The present Torquer system uses a variable chopper frequency to con-
trol the power flow at low speeds. Reduced frequency is required to
lessen the effect of undesired commutation power. Analyses of the EMT
system have revealed that a single frequency (1 KhZ) chopper mode of opera-
tion is desirable both from the standpoint of component efficiency and
cost reduction. Various techniques are possible for using a constant fre-
quency system such as auxiliary commutation or power pulse skipping coupled
with synchronized commutation command. The latter is an attractive scheme
since no additional power circuit elements are required. Fast turn-off
SCRs can be used with the EMT circuitry, allowing further reductions in the
size of commutation circuit components and affording increased system
control range (Section 6.5).
The ramp and pedestal control technique utilized on the breadboard
Torquer PCU*offers a means for an improved Speeder PCU. Other techniques
are suggested for improving the Speeder control (Section 7.4).
*power control unit
-------
Solid state switching of the Torquer PCU between the drive and re-
generation configuration has been demonstrated. The development of this
system proceeded in a series of steps from an all electromechanical relay
contactor system, to one in which SCRs and diodes are employed for current
direction control (Section 8.1).
Advanced Torquer PCU systems should employ several safety protection
and detection features including overcurrent and oscillation failure cir-
cuits along with thermal overload protection (Section 6.6). The advanced
Speeder system will provide primary protection by field sequencing. Faults
will be detected by oscillation failure, overcurrent and low battery volt-
age detectors (Section 7.5).
The EMT hybrid relaxes the driveability requirements placed on the
internal combustion engine and allows the engine to operate in a manner so
as to reduce emissions while retaining fuel economy.* Emissions of hydro-
carbons (HC) and carbon monoxide (CO) during the period immediately follow-
ing engine start-up account for a majority of the total emissions which
occur during the DHEW cycle test. A primary factor affecting ICE cold
start emissions is the rate of choke removal. Experiments indicated
that the hydrocarbon (HC) and carbon monoxide (CO) mass rates can be re-
duced 50% and 80-90%, respectively, from that of a conventional engine by
increased choke relief rates (Section 9.2).
In a hot ICE, CO and HC mass rates are relatively insensitive to air-
fuel ratios beyond stoichiometry (^18-19) and are directly related to
engine power level. Oxides of nitrogen (NOX) are quite sensitive to air-
fuel ratio and the NOX mass rates decrease rapidly with increased air-fuel
ratios in excess of stoichiometry. NOX mass rates also show a non-linear
increase with increased power level. Spark retard causes a slight decrease
in HC and CO levels and a strong decrease in NOX. As engine power level
is increased, leaner mixtures can be burnt without adversely affecting HC
and CO production (Section 9.2.3).
*"An Electromechanical Transmission for Hybrid Vehicle Power Trains -
Design and Dynamometer Testing", G.H. Gelb, N.A. Richardson, T.C. Wang,
and B. Berman, Society of Automotibe Engineers, Paper No. 710235,
January 1971.
-------
The 1975 standards for CO can be met with rapid choke removal and
lean carburetion. While the HC levels are reduced by leaning, the standards
are exceeded within a few minutes after engine start. A three-component
catalyst system was tested which shows some promise for NOX and HC control
during hot operation. The catalyst performs best near stoichiometry,
suggesting a fuel economy advantage over lean engine operation (Section 9.3),
Several actions are recommended in order to more fully explore the
potentialities of the EMT system:
• Design requirements for advanced Torquer and Speeder machines
should be developed into specific designs integrated with the
power conditioning equipment. Second generation hardware should
be procured and the improvements experimentally verified.
• A second phase of power conditioning equipment development should
be initiated. The new equipment should incorporate the circuit
features determined in this study. Emphasis should be placed on
utilizing lower cost power elements and developing packaging and
vehicle integration concepts.
• Work should be directed toward improving the gearing of the EMT
and incorporating the clutching and activation functions into the
gear box.
t Methods of controlling the prime mover power level in the various
operating modes should be investigated. Trade-offs and hardware
development of the activators and Torquer interfaces should be
conducted.
• Development of specialized ICE exhaust control approaches should
be undertaken including further experimental evaluation of catalyst
performance during the cold period. Fuel usage rates should be
established for various mixture ratios. Exhaust recirculation
coupled with a simple oxidizing catalyst should be evaluated in
terms of the unique manner of EMT-ICE operation.
• Further dynamometer test work should evaluate the dynamics of the
power train during drive-regenerate transfer, mode transition and
acceleration from standstill. The smoothness of system operation
should be defined and means explored, if necessary, to refine
system stability.
• Definition of major components should be made for a variety of
vehicle sizes ranging from urban vehicles through full-sized
passenger cars.
• The economy of hybrid operation should be evaluated and combined
with other system costs to establish the cost of EMT-hybrid owner-
ship.
-------
INTRODUCTION
TRW Systems has performed a study entitled "Analysis and Advanced
Design Study of an Electromechanical Transmission" under contract: numbor
EHSH-71-002 for the Air Pollution Control Office of the L'nvi roimienUtl Pro-
tection Agency. This report presents the results of work performed during
the period July 13, 1970 to February 13, 1971. It is concerned with the
eight (8) tasks of the contract as originally negotiated. A ninth task
associated with the study of the transient emission behavior of an internal com-
bustion engine during its cold period of operation was added to the con-
tract, extending the contract period to April 26, 1971.
This report deals with the design and data analysis of a heat engine-
battery hybrid propulsion system for vehicles. In particular, the work
focuses on the analysis and design of a transmission termed the Electro-
mechanical Transmission (EMT). The report is divided into nine (9)
sections structured in such a way that the reader is first acquainted wit.h
the basic features of the EMT. Later sections treat design tradeoffs
developed during the construction of breadboard, proof-of-principle equip-
ment and the testing of that equipment on a dynamometer. The remaining
sections are addressed to design changes which have or should be undertaken
to improve system performance.
Section 1 presents a short review of the basic operational modes of
the transmission developed by TRW Systems, the Electromechanical Trans-
mission, EMT. It describes and stresses the functional requirements placed
on the equipment by the hybrid system.
Section 2 discusses the traction system of the EMT and in particular
to various methods of controlling the Torquer (the traction device of the
EMT). A short discussion of candidate systems is presented leading to the
selection of a DC chopper using pulsewidth modulation. Various TRC*
philosophies and circuits are reviewed and their salient points discussed.
Finally, a specific TRC system and circuit are chosen and its operation and
critical components are described.
*time ratio controlled
-------
The generator system (Speeder system) is discussed in Section 3. A
review is made of various AC and DC system concepts in view of the unique
functional requirements of the Speeder system. A description of the bread-
board Speeder PCU is presented.
Section 4 describes the rotating machines used in the present system.
Manufacturers specifications and construction details are given.
The dynamometer testing of the EMT can be found in Section 5. In-
cluded are test descriptions, data reduction routines, and presentation of
tost data. Conclusions are drawn from the data useful for further design.
Problems encountered during the test program are discussed.
Section 6 contains the results of advanced analytical and design work
related to an improved Torquer and Torquer PCU. Included are recommended
methods of improving machine performance along with suggested machine rat-
ings in view of actual operating experiences. A computer-oriented analysis
or the Torque)- system is presented with trade-offs of major system para-
mo ters. The effects of frequency of chopper operation are indicated.
Reference is made to advanced logic control systems and the critical pro-
blems of Torquer-EMT interfaces and system protection are discussed.
Section 7 discusses the results of an advanced design study of the
Speeder and Speeder PCU. Included is an analysis of losses and stability
limitations in the present system. Failure modes are identified and
possible remedies are suggested.
Some advanced power train operational concepts are discussed in
Section 8. In particular a unique solid state drive-to-regeneration switch-
ing system for the Torquer is described. Certain hybrid power train start-
up concepts are suggested and related to possible emission control schemes.
Transmission override options are discussed.
Section 9 describes work conducted to establish the emission
characteristics of an internal combustion engine during start-up and
hot operation. Preliminary data on a three-component catalyst system
are presented. DHEW cycle emissions are projected.
-------
1.0 THE ELECTROMECHANICAL TRANSMISSION (EMT)
1.1 Description of Concept and Modes of Operation
The EMT 1s a particular form of a parallel-connected hybrid power train
which attempts to utilize heat engine energy as efficiently as possible
while minimizing the excursions of prime mover power and speed!- J
The EMT configuration and its basic power and information flow paths
are shown in Figures 1.1 and 1.2. Power produced by the engine is delivered
to the input (sun gear) of a planetary gear system. The planetary gear
system algebraically divides the input power between the planet carrier
and the ring gear in direct proportion to the speeds of those elements.
The speeds in turn are related through the system gear ratios to the in-
put shaft speed. In the present implementation the planet carrier gear
is connected to a generator--termed the Speeder—while the ring gear
drives the propeller shaft of the vehicle. A traction motor—termed the
Torquer--is coupled directly through a fixed gear ratio to the propeller
shaft.
In Mode One, as shown in Figure 1.1, the engine is run at constant
throttle, constant speed at a power setting consistent with the average
power required of the vehicle in urban operation. The constant torque
output of the engine is delivered to the Speeder and propeller shafts,
thus these two shafts receive a constant torque input in direct relation-
ship to the planetary gear ratios.
In order to maintain the engine speed constant it is necessary that
the Speeder produce an electromagnetic shaft torque which when reflected
through the gearing to the engine shaft is equal and opposite to that
produced by the engine. If the engine departs from the desired reference
speed, the differential information is used through the Speeder controller
to increase or decrease the Speeder torque and thus bring the engine into
control. The function of the Speeder as its name suggests is to control
the engine speed while the power train is in Mode One.
Because the engine-produced torque on the propeller shaft is constant,
the Torquer is used to augment or reduce the torque available to the wheels
for tractive effort. The Torquer therefore must be capable of operating
*Numbers in brackets refer to references in bibliography.
8
-------
INTERNAL
COMBUSTION
ENGINE
1
PLANETARY
GEAR TRAIN
ENGINE
SPEED ERROR
1
SPEEDER
(GENERATOR)
TORQUER
(MOTOR
GENERATOR)
SPEEDER POWER
CONDITION!
UNIT
TRACTION OR
REGENERATION
POWER
TORQUER POWER
CONDITIONING
UNIT
I
OPERATOR
COMMAND
BATTERY
FIGURE 1.1
EMT MODE ONE OPERATION
-------
wl T1
INTERNAL /> r\
COMBl
ENC
^
ENG
THRC
SIG
KTION AcADTDA IUI
-iKir \ \ GEAR TRAIN
?INE V V
* £ BRAK
£ & TOSF
r ~i
SPEEDER
L_ -I
1
'INE L_
JJI^ SPEEDER POWER
CONDITIONING
UNIT
L- -^ -1
y
BA
L
QJ T
o 'o
I i
E APPLIED
'EEDER SHAFT
TORQUER
(MOTOR OR
GENERATOR)
I
TORQUER POWER
CONDITIONING
UNIT
t
1
TTERY
^
r
/.j T
"T TT
/* /^~\
( i
i i
\ A \ A
\_y V-*
««.
i
i
DRIN
WH
i
i
i
^_ OPERATOR
^" COMMAND
|
1
/ING
EELS
i
ill
FIGURE 1.2
EMT MODE TWO OPERATION
-------
as both a motor and as a generator, and as its name suggests funct.inn f.<>
modulate the propeller shaft torque delivered to the wheels.
At a predetermined speed typical of the departure from urban opera-
tion, the transmission transfers into Mode Two as shown in Figure 1.2 and
allows the engine to operate at higher power levels consistent with high-
way cruise. In this mode a small brake is applied to the Speeder shaft,
stopping it and locking it out of the power flow. The engine is now
directly coupled to the rear wheels and its speed must change directly
with road speed. Throttle level of the engine is adjusted on the basis
of electrical power flow in the Torquer in such a way that the throttle
is always moved to a power setting where the engine takes over the total
propulsion requirement. If the Torquer power demand increases, the engine
throttle setting slowly increases until the engine output matches the new
steady road demand. If the Torquer demand decreases or goes into a gen-
erator manner of operation, the engine slowly throttles back.
Figure 1.3 presents an operational map of the EMT power train.
Assumed engine power levels, vehicle weight and transition speed from
Mode One to Mode Two operation are indicated. The envelope of probable
maximum road power demand during urban operation was developed from an
analysis of driving missions in Los Angeles traffic.12J Five regions of EMT
operation are indicated in the figure.
The first region is associated with the low acceleration events in
which the engine torque available directly to the road is equal or exceeds
the road torque requirement. In this area the excess engine power is
diverted to the battery with both Speeder and Torquer acting as generators.
In the second region engine power is equal or greater than the road
power demand, but it is not available at the proper torque level. The
Speeder accepts engine power, transfers it to the Torquer which delivers
the power to the road at the required torque level. As indicated, the
excess engine power is directed through the Speeder to the battery for
charging.
The third region is that in which the road power and torque demands
exceed the engine output. Here the battery is called upon to provide
11
-------
FIGURE 1.3
OPLRATIONAl MAP Of LMT POWCR 1RAIN
15 HP I NO IN I
40 MPH MODf FKANSIFION M'l I l>
3000 LB VEHICLE
54 HP
ENGINE CANNOT SUPPLY EITHER
ROAD POWER OR TORQUE;
TORQUER ACTS AS MOTOR
SPEEDER ACTS AS GENERATOR;
POWER DRAWN FROM BATTERY
ENGINE CAN SUPPLY
ALL ROAD POWER;
TORQUER ACTS AS MOTOR;
SPEEDER ACTS AS GENERATOR;
EXCESS POWER TO BATTERY
>ENGINE CAN SUPPLY ROAD TORQUE AND POWER;
vSPEEDER AND TORQUER ACT AS GENERATORS;
^EXCESS POWER TO BATTERY
ENGINE UNTHROTTLED;
SPEEDER LOCKED;
TORQUER SUPPLIES OR
REQUI RES BATTERY POWER
REGENERATIVE BRAKING
BOTH TORQUER AND SPEEDER
SUPPLY POWER TO BATTERY
Kfy^y,ff • /,-. •. • . v/xvX v
25 30
VEHICLE-MPH
35
40
45
50
55
-------
peaking power. The torque converter operation of the Speeder-Torquer
combination continues.
Above the 40 mph mode transition speed the transmission transfers
Mode Two and the Speeder is removed from the power loop. As previously
indicated, the power train still acts as a hybrid but the engine must
throttle up and adjust its speed to meet the higher average road power
and speed demands.
The fifth region is associated with all deceleration events. Because
of the dual manner of Torquer operation, regenerative (and/or dynamic)
braking is built into the system. Regenerative braking in either Mode
One or Mode Two can be used for battery charging.
1.2 System Functional Requirements
1.2.1 Speeder System
The Speeder System must perform two functions; to serve as a major
mechanical-to-electrical energy converter, delivering engine output to
the battery bus and as an engine load controller, varying its load so as
to maintain the engine under speed control. It must be capable of opera-
tion as a variable speed, quasi-constant torque generator, while deliver-
ing electrical power to a battery bus whose voltage can rapidly vary under
changing battery loads and charge states.
1.?.2 Torquer System
The Torquer System must serve as a mechanical-to-electrical (genera-
tor) and electrical-to-mechanical (motor) energy converter. It must be
capable of responding in a controlled, stable fashion to operator input
commands so as to modulate the torque at the driving wheels. It must
operate over a wide range of shaft torque levels while taking or deliver-
ing electrical power from a battery bus whose voltage can rapidly vary
under changing battery loads and charge states. In addition, the Torquer
System must provide primary traction power flow information to the prime
mover power control system so that the prime mover may be adjusted to
reflect Mode Two (high speed) operation.
13
-------
2.0 TORQUER SYSTEM
2.1 Review of Candidate Torquer Systems
The manner of torque and power control, cost and complexity of the
machine and controller are all dependent on the machine selected.
The electromechanical torque of all electric motors is proportional
to the integrated product of the radial magnetic field in the air gap and
the axial current in the armature (in space phase with the magnetic field).
The rating of an electric motor is limited by three factors:
• magnetic saturation of the iron circuit
t temperature rise due to internal losses
• mechanical stresses imposed on the rotating parts
Depending on its design, the horsepower/weight ratio, torque-current,
and torque-speed characteristics of the various types of traction motors
can be quite different. The method of speed control, regenerative or
dynamic braking, and the complexity of power conditioning may also vary
appreciably. In general, the series DC traction motor and the low fre-
quency AC series commutator motor have dominated the field of variable
speed electric drive systems because of their speed-torque characteristics
and simplicity of control. However, some recent experimental vehicles
have been constructed using induction motors which have used the simpler
rotating machines at the expense of sophisticated inverter power circuits.
Table 2.1 presents a list of the machines which could be used for the
Torquer. The table summarized important features of each machine as well
as its control equipment. While the AC synchronous machines do not have
a mechanical commutator as do most DC machines, they require additional
rotor position information for torque and speed control, making their
power conditioning systems somewhat complex and costly. Synchronous
machine systems do not have the simplicity of control as DC machines nor
the rugged, low cost construction of induction machines. In view of this
and since they have not found any favor in electrically propelled vehicles
to this time, they are not considered further.
14
-------
TABLE 2.1
CANDIDATE TORQUER SYSTEMS
Type of
Mflchi ie
I. DC Machine
a. shunt
b. series
c. compound
d. brushless
Starting
Torque
low-
medium
high
high
low-
medium
Commutator
or
Slip Ring
commutator
commutator
commutator
none
Relati ve
Rotor
Speeds
lowest
lowest
lowest
highest
Relati ve
Power
Densi tv
low
modest
modest
high
Can be Used
as
Generator?
yes
yes
yes
yes, may be
difficult
Relative
Machine
Cost
highest
highest
highest
medium
External
Power
Conditioning
DC chopper
DC chopper
DC chopper
DC-AC with
position
sensor
Relative
Complexi ty
of PCU
simplest
simplest
simplest
complex
Rel
Reli-
ef
ati ve
ability
PCU
highest
highest
highest
fair
II. AC Synchronous
a. conventional
j
i
b. permanent
magnet
c. solid
rotor
medium
low
low-
medi urn
slip ring
none
none
medium
highest
hi ghest
modest-
high
high
highest
yes
may be
difficult
yes
medium
un-
known
un-
known
DC-AC with
frequency
control &
position
sensor
most
complex
most
complex
most
complex
lowest
lowest
lowest
III. Induction Machine
a. wound rotor
b. squirrel cage
medium
medi urn
slip ring
none
medi urn
high
high
high
yes**
yes*
medium
lowest
DC-AC and
frequency
control ,
also re-
quire re-
active
power
SUDD! v
complex
complex
medi
medi
urn
urn
*Need external oower conditioning
**Need reactive oower s
-------
It should be noted that the average machine losses in a parallel
connected hybrid as the EMT are smaller than an equivalent weight vehicle
employing electric traction only. The ratings normally applied to motors
should be reduced in EMT application since the average propulsion power
(road load) is supplied by the prime mover and not the electrical motor.
2.1.1 DC Machine Systems
The simplest system which might be used for the Torquer would be a
reostat or contactor controlled DC motor. The literature contains a great
deal of information relative to these systems dating back to the earliest
days of vehicle traction. It is generally accepted that such systems have
the following disadvantages:
0 Reostat System
- poor efficiency at part load
- limited regenerative capability
• Contactors
- jerkiness of operation
- non-uniform utilization of the battery cells
At the present time the combination of power semiconductors with DC
motors and sliding contact commutators offers one of the simplest and
cheapest means of providing adequate starting torque, speed control, and
regenerative or dynamic braking.
The relationship between the torque, T, and armature current, I , of
a
a DC motor is:
T = 0.1174 x 10"8 pZa $ Ia Ib-ft = KA IA
where:
p = number of poles
Za = number of turns per armature circuit
4> = flux per pole (lines)
For the armature materials to be fully utilized, the magnetic materials
should be saturated and the armature conductors should carry a current
density consistent with the cooling capability of the machine. At very
16
-------
low speeds the electric loading may have to be reduced because of poor
ventilation. But, to a first approximation, the product of p and Z,K is
d d
constant for an armature of given dimensions. Therefore, the dimensions
of a DC armature and hence also of the whole machine, may be regarded as
boing functions of its rated torque.
The power output/input weight of DC machines increases rapidly with
the machine's rotational speed. However, the maximum speed of a conven-
tional DC machine is limited by mechanical commutator contact require-
ments and centrifugal forces on the armature windings. In an electric
vehicle traction machine power densities ranging from 0.06 to 0.2 hp/lb
are typical. It is only by going to brushless configurations that high
power densities are possible, approaching 1.0 hp/lb.
The brushes of DC machines are typically claimed as a major dis-
advantage of this type of machine. The brushes do wear in service and
must be replaced. The additional weight and volume that the brushes,
brush holders and commutator add could be used in an induction machine
for the production of torque and power. In terms of wear, brush life has
been cited in one case as a minor problem, with a useful life expectancy of
50,000 miles?-' With proper care to avoid pitting of the commutator, long
lifetimes can be expected. While the increased weight, volume and cost
of brushes and commutators are disadvantages they must be traded off
against the increased cost and complexity of induction machinery power
control systems.
2.1.2 Induction Motors
A polyphase induction motor consists of stator (primary) and rotor
(secondary) windings. When supplied from a voltage or current source, a
"rotating" magnetic field is set up by the primary winding currents which
induce currents on the secondary side. The secondary winding is closed
either by a short circuit in a squirrel-cage motor, or by an external
impedance in the wound-rotor motor. The interaction of the magnetic field
and secondary current produces an electromagnetic torque which is main-
tained as long as there is relative motion between the structure of the
primary and secondary windings. Since the secondary currents are created
by induction alone, no commutation or brushes are needed, and higher speed
17
-------
operation is possible. Advantages of this machine are its rugged and
simple construction and high horsepower/weight ratio. For these reasons
it has received some interest in recent times for vehicle use.
The torque-speed characteristic of an induction motor can be
expressed as
T =
where:
V = voltage input per phase
r, = primary resistance per phase
r~ = secondary resistance per phase referred to the primary
x = total equivalent leakage reactance per phase
n = rotational speed of the rotor
2f
n = — = synchronous speed (f is the source frequency, p is the
p number of poles)
n -n
s=-5- -slip
In the typical torque-speed characteristics of the induction motor,
there is a region where the motor is stable. This region is defined as
s < s < o where s is the maximum or breakdown torque. In this region,
s < s , a reduction in load causes the machine to speed-up. In practice
the point s = s is unstable and the maximum obtainable torque at rated
current is considerably less than the maximum speed. Using typical
values of machine parameters in the torque-slip equation it can be shown
that for a low slip (high efficiency) induction motor:
• the breakdown torque is less than three times the rated torque
« the line current at maximum torque is approximately four times
rated current
• the power factor at maximum torque is approximately 0.7.
These values compare unfavorably with a series DC motor. The series
machine is capable of producing three to four times the torque per
ampere as is the induction machine. Moreover, the DC motor does not
exhibit the stability problem of the induction motor.
18
-------
2.1.3 AC Motor Control Systems
The-speed of an induction motor can be controlled by one or a com-
bination o1 the following:
• varying the applied voltage, V
• changing the synchronous speed, nr> by varying the applied fre-
quency or changing the number of poles
• varying the resistance of the motor windings (wound rotor con-
struction).
The most satisfactory means of speed control is to vary both V and
f, yet maintain their ratio constant. When V/f = B = constant, the
machine will be able to operate with essentially constant flux density,
yet will be capable of producing almost constant torque throughout most
of the operating speed range. (At very low speeds the effect of stator
resistance becomes extremely important and higher terminal voltages must
be applied to develop constant torque.)
There are several control systems which have been developed for the
control of AC machines (synchronous and induction); they can be broadly
classified into:
• DC link inverters - those devices which accept DC power and invert
it to a controlled voltage and frequency
• Cycloinverters - those devices which transform AC power to a con-
trolled lower frequency and controlled voltage.
Since the latter devices are driven by an AC source they are not relevant
to battery powered systems.
There are two types of DC link inverters for AC motor control. They
are the pulse width modulating inverter and the variable voltage and
frequency inverter.
In the variable voltage, variable frequency inverter, the DC source
voltage is first chopped to the desired voltage level and then inverted
to the correct frequency. The pulse width modulated inverter performs
these two operations using a single set of components. While the former
system contains more components, the waveform presented to the AC machine by
the latter has a much lower harmonic content, offering better machine effi-
ciency.
19
-------
2.2 Selection of Torquer System
The selection of a system to use as the Torquer system involves an
assessment of both ability to perform required functions and the cost to
perform those functions. In terms of overall system cost, a judgment is
difficult. To make a rational decision involves a projection of the unit
costs of machines and solid state devices which are presently only manu-
factured in small quantities. However, in view of the present cost of
semi-conductor devices, it is reasonable to assume that while their cost
may fall significantly in the future, the large number required for AC
systems will still result in a total system cost higher than that of a DC
system using a more expensive machine. Coupled with the simpler control
system, higher reliability (fewer solid state components) and better
machine characteristics, a DC Torquer system was selected for the EMT.
2.3 Time Ratio Control (TRC) Systems
2.3.1 General Considerations of TRC
TRC is that method of control which employs a power switching device
(designated the power pass element) between a power source and a load.
The power pass element either allows full conduction of power or completely
blocks the power flow. The pass element alternates between the conduct-
ing and blocking states on a cyclical basis according to a controlling
signal. Power delivered to the load is controlled by adjusting the pass
element duty cycle. A variety of devices may be used as power pass
elements — contactors, SCR's (silicon controlled rectifiers) or transistors.
A more complete description of switching elements and their limitations
can be found in Section 2.4.
The basic elements of a TRC circuit are shown in Figure 2.1; they
include a power (voltage source), a power pass element (switch), and a
load.
The following general definitions apply to TRC control methods:
• chopping frequency, fc, the cyclical frequency at which the
power pass element changes from conduction to blocking state
t chopping period, T = j- = tQn+ tQff
20
-------
POWER PASS ELEMENT
OFF
o-
ON
LOAD
FIGURE 2.1
SIMPLIFIED TRC CIRCUIT
21
-------
on-time, t , the conduction time per cycle
off-time, t .., the blocking time per cycle
duty cycle, u, the ratio of on-time to chopping period; i.e.,
on
a =
T '
Figure 2.2 shows a circuit technique for voltage step-down (buck
control of a series wound machine). The switch, S, opens and closes in
a controlled fashion, giving rise to the current waveforms in the motor
and battery as shown in A . The diode is included to provide a path
for the armature current when S is opened. For the limiting condition
of very large series inductance, L, the armature current, I/\, will be as
in case B, and .. r
— a VR~ tn
I. =
A "a
Neglecting the resistance losses it can be shown that the average
(DC) terminal voltage of the machine, VT, is related to the
battery voltage, VR, as:
TT ' « VB
The circuitry included within the dashed lines of Figure 2.2 is termed
the power control unit, PCD, and may be regarded as providing DC voltage
transformer action as described by the above equation.
There are other power control functions to which TRC can be applied.
The Speeder system of the EiMT utilizes a boost; that is, voltage step-up
technique for transforming (boosting) the rectified voltage output of the
Speeder to battery bus voltage. The concept of intermittent power source
to load connection is again used, although in a different configuration.
The boost technique will be more fully described in Section 2.5.8.2.
22
-------
FIGURE 2.2
CIRCUIT TECHNIQUE FOR VOLTAGE STEP-DOWN
(BUCK) POWER CONDITIONING
z
UJ
0£.
ID
U
o:
a:
oc
U
<
CD
A. GENERAL CASE
UJ
D
<
on'
off-
Qi
a:
U
on
'off
TIME
-•-fc-TIME
B. LIMITING CASE
(LA VERY LARGE)
23
-------
2.3.2 TRC Techniques
2.3.2.1 Description of TRC Techniques
There are several control schedules which can be used to modulate
the TRC duty cycle, a. The methods include:
• constant frequency
- controlled on-time
- controlled off-time
e variable frequency
- constant on-time
- constant off-time
- variable on-time/variable off-time
Other methods such as pulse position modulation and pulse amplitude modu-
lation are not considered. Their relative complexity and cost of imple-
mentation do not appear to result in substantial system performance gains
over those enumerated above.
Constant Frequency TRC
Controlled on-time (Figure 2.3A) is that method of TRC in which the
conduction time of the power pass element is coincident with the duration
of the controlling timer or clock. This technique has the property that
when the control circuit is operated open loop and a particular on-time
is programmed, any errors in the clock frequency are absorbed, completely,
by the off-time. This method is characterized by the clock pulse simul-
taneously triggering the power pass element into conduction and synchron-
izing the variable time pulser, and the variable time pulse triggering
the cessation of the conduction of the power pass element.
Controlled off-time (Figure 2.3B) is the method of TRC in which the
blocking time of the power pass element is coincident with the duration
of the controlling timer. This technique has the property that when the
control circuit is operated open loop and a particular off-time is pro-
grammed, any errors in the clock frequency are absorbed, completely, by
the on-time. This method is characterized by the clock pulse simultaneously
triggering the power pass element into the blocking state and synchronizing
24
-------
a=25% a=50% a=75%
oFNFzuuuuui Jinnim nnnn
—H T h— -H T H- —~| T f—
CLQCK i i i i i i i i i i i i i i
PULSE
VARIABLE I I I I I I I I I I I I I I
TIME PULSE
FIGURE 2.3A
CONSTANT FREQUENCY-CONTROLLED ON-TIME
o = 25% o=50% o=75%
°FNF—JUUUUL nnmm nnnrn
—H T h— —H T h— —H T h—
CLOCK I I i | | I I I I i 1 j i i i
PULSE
VARIABLE I I I i I I I i I I I I i I I
TIME PULSE
FIGURE 2.3B
CONSTANT FREQUENCY-CONTROLLED OFF-TIME
25
-------
the variable time pulser, and the variable time pulse triggering the power
pass element into conduction.
In a system where a single timer is used for establishing the chop-
ping frequency and the commanded period (such as a ramp and pedestal
system as described in Section 7.3.1), the "on" time and "off" time are
interdependent and should be self-compensating in such a manner as to
maintain the desired duty cycle.
Variable Frequency TRC
Constant on-time TRC refers to that scheme in which the conduction
time per cycle remains unchanged over the entire range of duty cycles and
frequencies (Figure 2.4A).
Constant off-time TRC refers to that scheme in which the blocking
time per cycle of the power pass element remains unchanged over the entire
range of duty cycles and frequencies (Figure 2.4B).
Variable on-time/variable off-time is that type of TRC in which both
the conduction time per cycle and the blocking time per cycle of the power
pass element are adjusted simultaneously in order to vary the duty cycle.
Constant frequency operation is a special form of this control method.
2.3.2.2 Trade-off of TRC Methods
An analysis of the above control methods was carried out to determine
their relative merits in terms of: (1) the linearity between input
commands and output levels, i.e., how the machine terminal voltage, V,.,
varies with the control variable (f , t , etc.) and (2) the time required
for an input command to be realized as an output V,.. In the analyses it
was assumed that the TRC system would obey the idealized situation, i.e.,
'on
a = T :—r-
,f
on off
VT = <* VB
26
-------
a = 25% a = 50% a = 75%
n _ n J-LTLJI nnnn.
T
ON I I _ | _ I I I _ I I I I
PULSE
OFF _ | __ | _ | _ | I | _ | I I I
PULSE
FIGURE 2.4A
VARIABLE FREQUENCY- CONSTANT ON-TIME
a = 25% a = 50% a = 75%
op'HUULJLJ T-TLTLr U U LT
ON I I I I I I I |
PULSE
OFF I 1 I I I I I |
PULSE
FIGURE 2.48
VARIABLE FREQUENCY- CONSTANT OFF-TIME
27
-------
In each case the range of VT desired is between zero and VR as the duty
cycle varies between zero and one.
Figure 2.5 and 2.6 indicate the variation of V,. with the respective
control variables. In all cases, there is a linear relationship of out-
put to input and on this basis there is no reason for the selection of one
system over another. However, when one examines variable frequency opera-
tion in the light of the time required for a response to occur with a
change in command (Figure 2.7) the following observations can be made. In
order to obtain the full range of VT control in variable frequency opera-
tion it is necessary that the frequency range be from zero to infinity.
In the constant off-time mode, full VT requires 7 -»• °°; in the case of
_ i I fc
constant on-time zero V, requires ^ ->• ». In both cases, the requirement
1 Tc
of infinite period length would impose extreme slowness of system response
to variations in input commands. At the other end, the requirement of
j -*• 0 entails extremely high frequency operation. In view of the fre-
quency limitations of all candidate power pass elements, full range opera-
tion of the TRC would be limited.
There are secondary reasons for the selection of a constant frequency
system which will be described further in Section 6. However, in passing
it is important to note the simplification of magnetic and semiconductor
design with constant frequency operation. In turn, the ability to design
about one operating frequency will minimize filtering requirements, reduce
power peaks, losses, and component sizes and reduce potential stability problems.
2.4 Selection of Power Pass Elements Tor TRC
Three types of components are available for the power switching
functions of a TRC system; these are: electromechanical contactors,
transistors, and silicon controlled rectifiers, SCRs. A selection of
the appropriate device to use as the power pass element must be based on
consideration of:
• maximum current capability
• maximum voltage capability
• speed (switching time)
• maximum power capability
28
-------
V, = t f V0
T on c B
0
on
r
on f
CONSTANT FREQUENCY - VARIABLE t
on
VT=Kff)fcVE
0
0
roff
(o > t tr > 1-
\ off f
CONSTANT FREQUENCY - VARIABLE
FIGURE 2.5
off
29
-------
f ~foff
V =
\ f on / on
vr
roff
0
-------
VARIABLE FREQUENCY - CONSTANT ON-TIME
t - K0
on 2
Vr~M—VE
off on
Ko
VARIABLE FREQUENCY - CONSTANT OFF-TIME
foff K3
on
FIGURE 2.7
31
-------
• ease of turn-off
• switching and conduction losses
• gain, gate or coil requirements
2.4.1 Electromechanical Contactors
Maximum Current
Devices are available with current ratings so high that there is
essentially no current restriction in the present application. However in
service, the contactor must make and break high current levels repeatedly.
Contactors can be valuable switching elements if provisions are made to
insure the contactor is not operated while current in flowing through its
contacts. In this latter sense, they have been employed within the EMT
system (see Section 8.1).
Maximum Voltage
Devices are available with voltage ratings so high that there is
essentially no voltago restrictions in TRC systems.
Switching Time
Speed varies with the particular device selected but is typically
of the order of 100 milliseconds for heavy duty devices.
Ease of Turn-Off
Turn-off is accomplished by either energizing or deenergizing the
solenoid coil. An independent low level power source is required for
energizing.
Switching and Conduction Losses
Losses are associated with the resistance of the contact and the
solenoid coil. Although most losses occur in the solenoid coil, the
losses in the contacts must be reckoned with and depend on load condi-
tions and frequency of switching. Coil requirements depend on the size
of the contactor and its KVA rating, the larger the rating, the more coil
energizing power. For example, the contactor of the present Torquer system
uses a 25 W, 24 V coil to make or break contacts rated at 300 amperes.
32
-------
2.4.2 Transistors
Maximum Current
Currents up to 60 amperes can be handled in devices capable of block-
ing 300 volts in the forward direction.
Maximum Voltage
Transistors capable of blocking up to 1000 volts are available; how-
ever, their current capacity is limited to approximately 15 amperes.
Switching Time
The blocking-to-conduction time for transistors can be less than
two microseconds.
Ease of Turn-Off
Transistors are conventionally turned-off by removal of the base
drive. Turn-off time may be further reduced by application of a reverse
bias of one-to-three volts to the base-emitter junction.
Switching and Conduction Losses
There are three sources of losses in transistors:
t Losses during the rise and fall time of current flow. These
losses are due to the product of current flow times the forward
voltage drop in the drive during the switching period.
• Collector saturation losses equivalent to the steady state (con-
duction) voltage drop through the device.
• Base drive losses related to voltage drops in the base (control)
portion of the transistor.
The sum of the rise and fall time losses and the base drive losses
are frequently of the order of the collector saturation losses. In the
60 ampere transistor noted above, approximately 60 watts are dissipated
during the switching on and shutting off of the device. A similar power
loss is associated with its steady state operation.
Gain
Transistors typically exhibit a current gain of 25 in analog control
usage; however, when applied to saturated switching service as in a TRC
system, the gain is derated to 10. The lower gain is used to insure the
33
-------
transistor remains saturated (over-apply the base signal to insure? "a
closure" of the switch) and to insure that the transistor traverses it'.
analog region as rapidly as possible to approximate a switch more closely
and to reduce dissipation.
It is important to note that transistors cannot resist voltages in a
reverse direction and in circuits where a reverse bias is anticipated
blocking diodes must be provided in series. This requirement adds not
only additional complexity and sources of dissipation, but elevates the
base-to-ground voltage with increased susceptability to noise.
2.4.3 SCRs
Maximum Current and Voltage
SCRs are presently available which can carry 470 amperes, block volt-
ages up to 1200 volts, and retain a turn-off time of 15 microseconds.
While units are available which can block up to 1700 volts and carry 850
amperes, their turn-off time is greatly increased. For example, a device
with a 1200 volt, 1000 ampere rating will require approximately 150 micro-
seconds to turn-off. As will be demonstrated later, turn-off time impacts
heavily on the design and performance of a TRC system.
Switching Time
SCRs turn-on (reduce from an open-circuit state to a 5 volt forward
drop) virtually instantaneously. They turn-off (regain blocking capability)
in relatively longer periods. Turn-off times increase rapidly with current
capacity and temperature. The best units available today in power levels
appropriate for vehicle traction use have turn-off times in the neighbor-
hood of 10 to 20 microseconds.
Ease of Turn-Off
SCRs cannot be turned off solely by the removal of the gate signal
as in the case of a transistor; rather special commutation circuits or
AC techniques must be used in conjunction with the removal of the gate.
In any system the SCR is turned off by bringing the current flow
through the device to zero. This is done by removing the voltage source
or reverse biasing the voltage across the device. If provided by circuitry
(such as clamping a capacitor across it) it is called forced commutation.
34
-------
If provided by the power source such as the normal voltage reversal of an
AC supply it is called natural or line commutation.
Switching and Conduction Losses
SCR losses can be separated into five basic components as follows:
• gate trigger losses
• turn-on losses
• forward conduction losses
• commutation losses
• turn-off losses
Gate losses are low typically 300 mw due to the very high switching gain.
The effects of gate losses may be reduced even further by applying low
duty cycle pulse commands as opposed to steady state commands.
Turn-on losses are significant during approximately the first 40|is
following application of the trigger pulse. During this time the device
passes through a highly dissipative analog region, where its forward
voltage drop remains high. The steady state conduction losses are equal
to the product of the SCR saturation (steady state) voltage-approximately
2.5V and the load current.
Commutation losses and turn-off losses are similar to turn-on losses.
While commutation losses in the SCR itself are of the same order of mag-
nitude as the turn-on losses, the circuitry which provides the energy for
turn-off has losses associated within itself and must be included as a
part of a total SCR-based TRC system.
2.4.4 Summary of Candidate TRC Power Pass Elements
It is apparent from the above considerations that electromechanical
devices are too slow for use in an advanced TRC system. In view of
the total number of switching cycles anticipated for vehicle use and the
limited service capability of such devices for high power switching
they are rejected.
35
-------
Transistors, while providing the necessary switching speed and "arc-
less" circuit breaking, are not currently available in a power handling
capacity sufficient for vehicle application with paralleling. Paral-
leling techniques introduce additional problems of base drive synchronira-
tion
-------
Advantages and disadvantages are cited for each circuit and thus a
note of explanation is required so that some ranking may be established.
Circuits which require charging the commutation capacitor through the
motor have two drawbacks. First, the rate of voltage build-up and the
final voltage level of the capacitor will depend to some degree on the
state of the motor back-EMF. In these situations, lockout protection
(circuits which detect the potential inability of the commutation capac-
itor to provide enough energy to turn the power SCR off and thus inhibit
the gating on of that SCR) may have to be included. Second, if charging
power flows through the motor, there is a certain minimum amount of energy
delivered per cycle to the motor through the commutation circuit even if
the power SCR is not turned on. Low speed traction control may be dif-
ficult.
Circuits which discharge the commutation capacitor through the motor
have an advantage and a disadvantage. The advantage lies in the more
complete utilization of commutation power in that not only does it turn
off the power SCR, it also provides tractive effort. On the other hand,
the commutation discharge energy per cycle presents a low speed control
problem as suggested above.
2.5.1 Circuit I (Figure 2.8)
Operation
Initially capacitor C, is discharged. SCRs 3, 1 and 2 are fired in that
sequence. When SCR 3 is turned on LI and C^ form an oscillating circuit.
The voltage across C, oscillates for 180° reverse biasing SCR., and turn-
ing it off. C-j is left charged to approximately twice-battery voltage in
a direction opposite that indicated in the figure. SCR, is then fired,
delivering a power pulse to the motor load. SCR, is commutated off by
turning SCRp on thus clamping a reversed voltage across SCR,. C, dis-
charges into SCR, and the load. As C, discharges, the current through
SCR? goes to zero, commutating that device off. The PCU is then ready for
the next cycle.
37
-------
U)
00
\
SW
SCR
"\
— BATTERY
./
FIGURE 2.8
CIRCUIT I
-------
Advantages
• The charging current for the commutating capacitor does not flow
through the motor load.
• Commutating energy discharges through the load.
t The commutation capacitor charges independently of the state or
duty cycle (a) of the power SCR, SCR], provided SCR2 is not turned
on before the charging oscillation is completed.
Disadvantages
• A specific SCR firing sequence is required, as well as synchroni-
zation of all gating signals.
• Three SCRs are needed.
• In the event SCR? fails to turn on, SCR] cannot be commutated off
and the circuit is endangered.
t If SCR2 is triggered before the charging oscillation is completed
SCR] cannot be turned off.
2.5.2 Circuit II (Figure 2.9)
Operation
Initially the commutating capacitor, C], is discharged. When the
power contactor, SW] is closed L] and C] form an oscillating circuit and
oscillate for 180 , leaving C] charged approximately to twice battery
voltage in the polarity indicated in the figure. The power
SCR, SCR] is then fired, delivering a power pulse to the motor. In order
to turn SCR] off, SCR£ is turned on, discharging C] into 1-2 and turning
SCR] off. \-2 and C] oscillate 180°, charging C] to an opposite polarity
and commutating SCR2 off. (L] must be much larger than l_2 to insure l_2
and C] form the major oscillatory circuit.) After SCR2 turns off, L] and
C] again form an oscillatory loop, charging C] for the next cycle.
Advantages
• The charging and discharging current for C-j does not flow through
the motor
• The charging of Cj is independent of the state or duty cycle of
the power SCR, SCR]
• SCR] and SCR2 need no starting sequence relative to each other but
must be synchronized.
39
-------
SW1
CURRENT
SENSE
FIGURE 2.9
CIRCUIT II
-------
Disadvantages
t If the oscillation of L« and C, does not commutate SCR^ off, or if
SCR2 assumes a permanent conduction state, the system is locked in
an unsafe mode with the battery connected across L-j and l_2, and
unlimited current in SCR2-
t Commutating capacitor energy does not flow through the motor.
• 1-2 is an additional loss member placed in series with the load
during the power pulse period when SCR-j is on.
2.5.3 Circuit III (Figure 2.10)
Operation
When SW] is closed, the commutating capacitor, C-j, is charged in the
polarity indicated via the saturable reactor SR-j, diode DR2 and the motor.
The path through l_2 and DR4 is provided in the event the effective imped-
ance of the load is too large to allow effective capacitor charging. SCR]
is fired to deliver the power pulse to the motor load. To turn SCR^ off,
commutating SCR2 is fired causing L-j and C] to oscillate through 180°,
charging C-j to a polarity opposite to that indicated in the figure. C] is
charged to a voltage equal to the battery. The saturable reactor is
driven into a saturated state at the negative peak of the 180° oscilla-
tion allowing C] to discharge through SR-j, DR2 and the load. SCR1 and SCR2
are commutated off and C] goes to its initial condition ready for the next
cycle.
Advantages
t The commutating capacitor charges completely, independent of the
state or duty cycle of SCR-j.
t C-] discharge energy flows through the load.
t No starting sequence is required for SCR2 relative to SCR].
Disadvantage
t Capacitor charging current flows through the load.
41
-------
ro
FIGURE 2.10
CIRCUIT III
-------
2.5.4 Circuit IV (Figure 2.11)
Operation
After SW] is closed, SCR] is fired, delivering a power pulse to the
motor and inducing a voltage in the secondary of saturable transformer T]
with terminals 1 and 3 positive, thereby charging C] in the polarity in-
dicated. SCR2 then fires discharging C] into the load and commutating
SCR-) and itself off. The cycle is then ready to be repeated.
Advantages
• Commutating capacitor discharges through the load.
t No firing sequence is required with respect to SCR-i and SCR2-
Disadvantages
t The- commutating capacitor charging current does flow through the
load.
t The voltage to which C] will charge is a function of the duty
cycle, the state of conduction of SCR] and the load. However,
depending on the minimum duty cycle and minimum load anticipated,
it may be possible to consider the charging of Ci independent of
the load and duty cycle of SCR,.
• If T] does not charge C] sufficiently, or if for some other reason
SCR] fails to commutate off, the system is locked in an unsafe
condition and the battery, motor and SCR] may be damaged.
• This circuit must operate with a variable frequency control system.
2.5.5 Circuit V (Figure 2.12)
Operation
SCR] is turned on delivering a power pulse to the load and charging
the commutation capacitor Ci to battery voltage at a polarity opposite to
that indicated in the figure. SCR-, is turned off by turning on SCRo,
clamping C, across SCR-,. SCR2 remains in the conducting state allowing
C] to charge through both the load and R^ to the polarity indicated. When
SCR-, is subsequently fired, SCR~ is commutated off by having C, clamped
across it.
Advantages
• There are no important advantages to this circuit
43
-------
o
SW1
^
.=_ BATTERY
D2l
D, -i
^ 1 t.
CURRENT
SENSE
r~
Ri
L
LA
IT
(
F
1
"1
»
4 ||
4 ||
s
)
MOTOR
CU/ 9 C\A/ T D
v JVV i jVVOIxf'
^^^B^ ^K ^~^^^^^i
5 ^±^^~~°"^FC
~
SCR 3
4
3
± SCR 2 SCR 1
FIGURE 2.11
CIRCUIT IV
-------
tn
FIGURE 2.12
CIRCUIT V
-------
Disadvantages
• The voltage to which C] will charge is a function of the duty cycle
and the state of conduction of SCR^ and SCR2. Although L-| is in-
cluded to provide resonant charging of C], commutation may still
be a problem at very short duty cycles.
« Capacitor charging current flows through the load.
o The commutation circuit includes major resistive components.
• The upper limit of chopping frequency is limited by R2 and C-j.
2.5.6 Circuit VI (Figure 2.13)
Operation
After SW^ is closed, the following starting sequence must be
followed. SCR2 turns on, charging C-| through the load to a polarity
opposite that indicated in the figure. When C-j is fully charged, current
in SCR2 ceases, turning SCR2 off. SCR1 is then fired, delivering a
power pulse to the motor and establishing an oscillatory circuit between
C] and L-|. After the current in the commutation circuit oscillates for
180°, C] is charged to battery voltage in the polarity indicated. To
conclude the power pulse, SCR2 is fired, discharging C-| across SCR-j corn-
mutating it off. C, continues to charge through the motor to an opposite
polarity.
Advantage
9 Capacitor energy is discharged through the motor.
Disadvantages
e The commutating capacitor charges through the motor.
e The voltage to which C] will charge is a function of the load, the
duty cycle and the state of conduction of SCR-j.
» A starting circuit is required; if a proper sequence is not
observed SCR-, cannot be turned off.
2.5.7 Circuits I through VI are compared in Table 2.2 in terms of the
major power handling components and their operational requirements. The
total number of critical power components includes all SCRs, diodes, chokes,
reactors and transformers which handle load or commutation power.
46
-------
SCR
i SW1
6 (
k
P
=_ BATTERY
•
,- ..n-_. - <
M*
•^»
SCR 2
>2>
~
r
1 "3
k.
F
J
R C
mi
FIGURE 2.13
CIRCUIT V!
vw
CURRENT
SENSE
RA
LA||!
II'
(
4
*
»
»
}
1
LF
1
>
J
= SW3 IR
0 F
-------
TABLE 2.2
COMPARISON OF CANDIDATE DRIVE CIRCUITS
Circuit
I
I
II
III
IV
V
VI
No. of
SCRs
3
2
2
3
2
2
No. of
Commut
SCRs
2
1
1
2
1
1
No. of
Power
Diodes
2
2
3
2
3
2
No. of
Chokes
1
2
2
0
1
1
No. of
Satur-
able
Reactors
0
0
1
1
0
0
Starting
Sequence
Required
yes
no
no
no
no
yes
No. of
Semiconductors
in Series
with Load
1
1
1
1
2
1
Auxiliary
Capacitor
Charging
yes
yes
yes
no
no
no
No. of
Critical
Power
Components
7
8
11
7
10
(counting
resistors)
6
-e*
00
-------
2.5.8 Selection of Torquer PCU Circuits
The circuits described in the previous sections have been used for
traction drive systems where only motor operation and in some cases
dynamic braking has been desired. The functional requirement of the
Torquer system requires that the Torquer be equally capable of providing
efficient and controllable regenerative operation. The following section
describes considerations given to this requirement and circuitry to per-
form those functions.
2.5.8.1 Single vs. Dual Controllers
Two techniques are available for providing drive and regeneration
control. One method is to utilize two independent power controllers as
shown in Figure 2.14. The second method is to employ a single controller
which can be reconfigured by appropriate switching to perform in both modes,
The two controller system offers an advantage in that the transfer
from drive to regeneration and vice versa can be made smoothly and with
vanishingly small dead time for the transfer. However, it is obvious
that such a system will contain nearly twice the major power processing
elements as a single controller. On the basis of a greatly lowered part
count and lower cost, methods of utilizing a single controller concept
were explored.
2.5.8.2 Consideration of Step-Down and Boost vs. Boost Power Processing
Systems for Regenerative
While functioning as a generator, the Torquer must be capable of
delivering power to the battery over a wide range of machine speeds and
battery voltages. Two concepts for power processing and control are
possible; step-down and boost and straight boost. While the discussion
presented here is within the context of Torquer PCU trade-offs, it will
also serve as a basis for trade-off decisions concernfng the choice of
a Speeder PCU system (see Section 3.3). The basic elements of both
circuits are shown in Figure 2.15.
In the step-down and boost system, switch S-j is opened and closed
according to the TRC principles previously outlined. When the switch is
closed, the generator EMF, Vg, is impressed across inductor L. When S]
49
-------
tn
O
(BOOST)
r
i
i
i
] TORC
1
1
1
1
1 -
«
4
n
3UER
4
<
4
1
[
> '
(BOOST)
VOLTAGE
SENSE
(BOOST) :
»
>
r vi
M5 (BUCK)
r
BUCK
ft* rONTfJOl
(DRIVE)
BATTERY JI
•v
MODE "OR" * xAorM:
rATp ^ — MODE
r OAIt SIGNAL
\ .«.-..
ir
BOOST
CONTROL
(REGENERATE)
" (BOOST) ^
CURRENT
SENSE
H-
(BUCK)
FIGURE 2.14
BUCK-BOOST TORQUER PCU WITH DUAL CONTROLLERS
-------
T
! I
_ «*
•i
j
STEP-DOWN AND BOOST PCU
•M-
7
BOOST PCU
FIGURE 2.15
COMPARISON OF TRC GENERATOR CONTROL TECHNIQUES
51
-------
is opened, the inductor voltage, L -rr must equal VB> and the inductor dis-
charges through the diode into the battery. Note that the voltage of the
generator is unconstrained relative to the battery voltage, that is V can
be greater or less than Vg. On the other hand, the current through the
generator is discontinuous.
In the boost system, switch S-j is also opened and closed. When it is
closed, the
-------
While consideration of the two systems on the basis of design flexi-
bility might favor the step-down and boost in that it allows use of a low
voltage battery (small numbers of cells) and a modest to high voltage
generator; the table shows that such a design attempt results in a greater
difference in the peak and rms current. As shown in the numerical examples
in the appendix the effects are:
• For equal torque capabilities, the losses in machine and chopper
of the step-down and boost system are higher.
• The inductor (either an external or series winding) will have to
be rated for higher currents to prevent saturation.
• The battery will operate at higher peak currents, increasing over-
voltage and gassing problems.
• All conductors, contactors and solid state devices will have to be
rated to carry higher peak and rms currents and dissipate more losses,
2.5.8.3 Adaption of Candidate Circuits to Torquer PCU Use
Circuits I through VI are shown only in the drive condition and thus
must be modified for regenerative operation along guidelines stated in
2.5.8.1 and 2.5.8.2. Circuits I and II contain commutating capacitor
circuits which are in parallel with the load circuit. Rearrangement of
these circuits to fulfill the boost configuration would require additional
circuit elements to intermittently break the commutation circuit path to
ground. The additional circuitry would introduce additional cost and com-
plexity factors.
Circuit IVcontains a commutation circuit in series with the load and
there does not raise the same objections of circuits I and II with regard
to the complexity of rearrangement for regeneration. However, the addi-
tional SCR required (Table 2.2), added to the disadvantages listed for the
circuit suggest that it is not a suitable circuit when compared to the re-
maining circuits III and VI. Circuit V has no advantage in this application.
Circuits III and VI are redrawn in Figures 2.16 and 2.17 with addi-
tional switches (SW4) so as to be capable of both drive and regeneration.
A more detailed qualitative and analytical comparison has been made of the
two circuits in order to arrive at a final decision.
53
-------
f4
1
D5
1 \4 i
" H '
O J
SW 1
I
>;
r
C
— BATTERY
•
^
^
C
M»
>
iv
>—
q
r
^
) « • - <
^»
SW"4 "1
SCR 1
C, SCR 2 ;
1 r
i^ rr^nr\ ,
*V
^v
r
> .
h
3
.D2
SR
1h
4
<
^
CURRENT
SENSE
r
RA
i (
k 1
F ^j
R <
MOTOR
"H
5
»
H
J)
L^l2 p1^
f LF
j
j
S SW3Rj
F
-------
>- 1 d
SW~4 (
o
SW1
SCRl
SCR 2
.=. BATTERY
CURRENT
SENSE
FIGURE 2.17
CIRCUIT Ml MODIFIED FOR REGENERATION
-------
Figure 2.18 shows the idealized current and voltage waveforms
impressed on critical components of circuit VI. SCR, serves a dual
function in the circuit; as well as conducting the power pulses to and
from the motor, it also serves to complete the oscillation path for C^
and L^. The current in SCR-j is therefore comprised of two components — the
motor current, IM, and the commutation capacitor charging current. The ideal
ized peak current in SCR, can be shown to be:
_
1DV cro = IM + •• where A,/C, is the characteristic impedance
rl\ oLK, IM f, ,,, II
I 1' 1
of the oscillating circuit. In a like manner, the average current in SCR,
can be shown to be:
2V C It
VB L1 !M *N
SCR] T T
The first term is the current contribution from the discharge of the
capacitor. Since the capacitor charging component exists approximately
for the period TT/(JC and the motor component for time t.., the RMS current
in SCR, is:
When SCR- conducts it must supply full load current to the motor plus
a current component equal to full load current to commutate SCR, off. The
peak current in SCRp is therefore equal to:
=21*
SCR2 £ *M
current must flow for the period t , the turn-off time of SCR,
Allowing a safety factor of 1.5 the average current in SCRp is equal to
*It must be noted that this peak exists for less than 10% of the turn-off
time t . Thus, for practical purposes, IpK SCR is equal to I...
56
-------
FIGURE 2.18
WAVEFORMS OF MODIFIED CIRCUIT VI
MOTOR
TERMINAL
VOLTAGE
BATTERY
VOLTAGE
'SCR i
—|i
'M
d
M
M
'SCR 2
m
57
-------
]-5xIMx2tc
SCR2 " T
Also:
I
RMS SCR2
(The transfer of load current from SCR, to SCR9 is done very rapidly and
ri T
SCR9 is subjected to a high rate of current increase (JT-). In those uses
H T
where -jr is large, local "hot spots" may occur in the SCR semiconductor
junction due to a lack of current uniformity. Turn-on switching dissipa-
tion may lead to an excessive temperature rise and premature SCR failure.)
The waveforms associated with the modified circuit of Figure 2.16 are
shown in Figure 2.19. When SCR, conducts, its current is the same as the
motor current, I... Hence,
DK cpp ~ M
rl\ oLK-i 1*1
Since SCR, conducts for the period t.,;
and
I
SCR]
RMS SCR1
When SCR« conducts, it permits C, to resonate with L, through 180°.
Since C, is charged to Vg at the beginning of this oscillation and the
circuit presents a characteristic impedance equal to
'PK SCR2 "
The time of this 180° oscillation is nA-j/C^. Since this occurs once
each power cycle, the RMS current in SCR2 is :
58
-------
2V,
MOTOR
TERMINAL v
VOLTAGE
B
B
BATTERY
VOLTAGE
FIGURE 2.ly
WAVEFORMS OF MODIFIED CIRCUIT III
'SCR i
'M
o
u
i
M
'M
SCR 2
r\ TV
VD A/L/c
A
—|
59
-------
SCR
Immediately prior to SCR2 conducting, C-| is charged to Vg . At the
conclusion of the commutation function C-j is charged to VB in an
opposite polarity. C, therefore traverses a voltage change equal to 2Vg.
The average current in SCR2 is therefore equal to
2V C
B Ll
SCR2
The current waveform in SCR2 is sinosoidal and thus the -rr impressed on it
is lower than that on the commutating SCR of the previous circuit.
Table 2.4 compares the component count and major operational features
of the two circuits. The total parts of each circuit are greater than
Table 2.2 to include the incorporation of switch SW., the switching element
required to transfer the circuit between the drive and regenerative con-
figurations. The count for circuit VI does not show the lockout logic
circuit required to insure safe start-up.
The modified form of circuit III shown in Figure 2.16 war, chosen for
hardware development over the modified form of circuit VI for the follow-
ing reasons:
• The commutating capacitor C] of circuit III charges completely
independent of the state of the load.
• There is no starting sequence required for circuit III, hence no
additional lockout circuits are needed.
t The peak and RMS currents of the power pass SCR, SCR], and the
commutating SCR, SCR2 are lower in circuit III.
dl
* ™e dtf stresses in SCR9 °^ circuit III are lower.
2.6 Description of EMT Torquer System
The Torquer system is shown in the block diagram of Figure 2.20. The
major components of the system are:
• Battery
• EMT Master Control
60
-------
TABLE 2. 4 CANDIDATE TORQUER PCU CIRCUITS
No. of No. of No. of Starting No. of Auxiliary No. of
No. of Com. Power No. of Saturable Sequence Semicond. Capacitor Critical
Circuit SCR's SCR, Diodes Chokes Reactors Required in series Charging Components
with load
Ill Modi-
fied
VI Modi-
fied
2
2
1
1
4
3
2
1
1
0
No 1
Yes 1
Yes
No
12
8
-------
FIGURE 2.20
EMT BLOCK DIAGRAM OF TORQUER ELECTRICAL SUBSYSTEM
en
EMT
MASTER
CONTROL
THROTTLE
TIME RATIO
CONTROLLER
AND CLOCK
TRANSFER
LOGIC
CURRENT
TRANSDUCER
DRIVER (PEDAL) CONTROL
CURRENT SIGNAL
OUTPUT
SHAFT
TACHOMETER
VEHICLE SPEED
-------
0 Pedal Control
• SCR Power Stage
t Buck/Boost Module
0 Field Logic
Torquer PCU
• Time Ratio Control and Clock
t Transfer Logic
• Torquer
2.6.1 Battery
The battery chosen for the EMT system is a 200-volt, lead-acid,
secondary battery. This voltage was chosen for the following reasons:
• A lower voltage would require larger current carrying capabilities
in all major circuits.
« A higher voltage, while lessening the current carrying capability
requirements would mean--
- more cells in series with increased reliability and maintenance
problems
- higher voltage rated semiconductors
- costlier insulation and manufacturing procedures.
2.6.2 EMT Master Control
The master control serves as a focal point for EMT operation in that
it provides the link between engine and electric subsystems as well as the
mechanism for Mode One-Mode Two transfer. The control receives vehicle
speed information and transfers control of the engine throttle from the
fixed value of Mode One operation to the variable level of Mode Two. In
Mode Two, it processes Torquer current signals into position signals for
the engine throttle servo system allowing the engine to slowly increase or
decrease power as a slave to the electrical traction system. It also pro-
vides the control and actuation signals for applying the brake to the
Speeder shaft during mode transfer.
63
-------
2.6.3 Pedal Control
The pedal control is the means for operator torque input into the EMT-
Torquer system. It consists of a pedal-driven potentiometer and switch
system. Pedal full-off generates a control signal which results in maxi-
mum Torquer regeneration. Increased pedal deflection reduces the regenera-
tion torque command. At approximately 50% of the full pedal travel, a
switch on the pedal is activated, signaling the transfer logic to establish
the buck module configuration and the drive polarity of the Torquer. In-
creasing pedal deflection commands increasing Torquer drive torque.
2.6.4 SCR Power Stage
The power stage consists of the power pass SCR and associated commuta-
tion circuitry of the modified version of circuit III. This stage processes
the power flow between the Torquer and battery according to a duty factor
and chopper frequency determined by the time ratio controller and clock
circuit.
2.6.5 Buck/Boost Module
The buck/boost module establishes the functional state of the PCU and
the wiring of the power stage. One position of a switch connects the module
for buck operation and the other position connects it for boost operation.
2.6.6 Field Logic
The field logic controls the polarity of the motor field according to
the "buck" or "boost" requirements.
2.6.7 Time Ratio Controller and Clock
The time ratio controller compares the value of the pedal torque
command with the actual Torquer torque (measured by the Torquer armature
current) and generates an error signal which causes the duty cycle (a) of
the power stage to change. The clock establishes the operating frequency
of the chopper system.
2.6.8 Transfer Logic
The transfer logic receives direction from the pedal control and
directs the Torquer PCU into the buck (drive) or boost (regeneration) con-
figuration.
64
-------
3.0 SPEEDER SYSTEM
3.1 System Requirements and Selection Criteria
The functional requirements of the Speeder system were described in
Section 1. They are summarized below:
• To regulate the engine speed during Mode One operation by providing
a controlled modulation of the electromagnetic torque reflected to
the engine shaft.
• To deliver power to the battery bus over a wide variety of Speeder
speeds and battery voltages.
These requirements must be satisfied by a system which offers low cost,
weight and size, with high reliability and efficiency.
3.2 Selection of the Speeder
There are various candidate generators which differ in construction
and control method details. Table 3.1 presents a qualitative assessment of
various generators which could be used as the Speeder. Only two specific
types of electric generators are considered suitable for Speeder use: con-
ventional DC generators and conventional synchronous generators. The con-
ventional DC machine is attractive because the electrical load is DC.
Conventional synchronous generators are also considered suitable for
this application. The synchronous generator (alternator) has been
completely accepted by the automobile industry over the DC generator. Even
though additional rectifiers are required for the DC load, AC synchronous
machines offer the advantages of lower cost, lighter weight and freedom of
commutation problems. In the case of the Torquer, the use of an AC machine
was complicated by the need for positional and/or frequency information to
be supplied to the PCU for torque control. If an AC machine is used solely
as a generator, these requirements are not present. In view of these con-
siderations, the conventional synchronous generator was adopted as the
Speeder.
3.3 Rational of Speeder Operating Mode and PCU Selection
In order to control the torque of a simple alternator one can control
the field,armature current or both. It can be shown that for constant
torque and voltage output, the field current must vary inversely with the
65
-------
TABLE 3. 1 CANDIDATE SPEEDER MACHINES
Type of Generator
Electromagnetic
Torque control-
led key-
Mechanical Relative Relative Generator
Commuta- Slip Windings Rotor Power Relative Stability
tor Ring on Rotor Speed Density Cost Problems
I. D. C. Generator
a. Conventional
b. Brushless D. C.
II. A. C. Synchronous
a. Conventional
b. Permanent
Ma gnet
c. Solid Rotor
field flux and
armature current
field flux and
armature current
field flUx k arma-
ture current
armature current
field flux and
armature current
Yes
No
No
No
No
No
No
Yes
No
No
Yes lowest low to high
medium
Yes high
No highest highest very
high
No
medium very Unknown
high at
present
Yes medium high medium Yes
No high medium high Yes
Yes
III. Induction Generator
a. Wound Rotor
stator current
and rotor
resistance
No
Yes Yes medium high
medium Yes
b. Squirrel cage
stator current
No
No
Yes hi ah
hi ah
lav
Yes
-------
machine speed. Under these conditions the field current becomes very
large at low speed and to avoid driving the field into deep saturation,
the machine must be over-designed with excessively large pole areas. At
high speeds, the armature current becomes large compared with the fiold
current and the machine must be designed to withstand armature reaction.
Electric rotating machinery is most completely utilized when it is oper-
ated with its magnetic elements just saturated and its copper conductors
continually carrying their design current. Therefore, the machine con-
trol concept which promises lightest weight and lowest cost is the one
which utilizes the machine at high speed, with a saturated field excita-
tion and constant armature current. Such a concept will produce a con-
stant torque machine having a voltage output approximately proportional
to the machine's speed. It is evident that a simple power conditioning
and control system is needed so that the variable output voltage of the
machine can be transformed to battery voltage in order to deliver machine
power to the battery.
A simple method of power conditioning is to tap change at the battery,
i.e., connecting the output of the generator to various voltage points
of the battery as the generator voltage varies. This method is slow act-
ing and leads to unbalanced battery charging.
3.4 Speeder PCU
The more attractive way of controlling the Speeder power is through
the use of a boost type controller utilizing the TRC techniques of Section
2.5.8.2. The type of PCU which was developed is designated as a voltage
boost type power processor with constant current capability. A schematic
of the Speeder PCU is shown in Figure 3.1. It is very similar to the
Torquer PCU in that it utilizes the same commutation techniques. However,
the PCU does not require mode transfer modules or logic. It does contain
additional elements; a full wave rectifier and energy storage inductor.
The latter is necessary since the armature of the alternator cannot be
used for energy storage as was the series field winding of the Torquer.
The commands for turning on the power SCR, SCR], and the commutating
SCR, SCRC, arise from a logic control system which compares the actual
engine speed to a reference speed and adjusts the duration of power SCR
67
-------
00
/
/^
^J
5PFFHFR \
GENERATOR
t| | M <
™ T ™
RECTIFIER
ENERGY
STORAGE INDUCTOR
1
s
L
iSCR]«
7
u
r*
SCR
c ,
X
SCR POWER SWITCH
BLOCKING
RECTIFIER
"1
, l
1
- 1
1 ^vwv f^R
1 kJ -^=
PRECHARGE
. I CIRCUIT
1
1
, l
> l
_J
k -fttt
TO
TORQUER
BATTERY
FIGURE 3.1
-------
conduction time. This in turn, determines the energy stored in the in-
ductor and the load reflected back to the engine.
Two additional control features are presently in use within the
Speeder PCU. The first is a torque lockout feature which turns the
Speeder field off when the engine speed is below 1300 rpm. This allows
the engine to be cranked and come up to power during system start-ups.
The second circuit feature is field control circuit which adjusts the
Speeder field to insure the Speeder output voltage does not exceed battery
bus voltage. In such a situation the Speeder PCU would lose control be-
cause the source current would flow into the battery unimpeded. This method
of voltage control is at odds with previous arguments, in favor of constant
excitation operation. This area is further discussed in Section 8.
69
-------
4.0 ROTATING MACHINES
4.1 Torqtier Selection
In Section 2.1.4 a DC machine was selected for the Torquer system.
The circuit trade-offs of Section 2.4 assumed the use of a series wound
motor. While there was no reason given in that section, a comparison of
shunt and series motors shows that series motors are better suited to the
conditions of the Torquer because of the following factors:
• The series motor develops a large torque at low speeds, ideal for
traction purposes.
• The series motor possesses a high free-running (or light load)
speed.
« The field of the series machine is only utilized when the machine
is producing torque; parasitic losses are avoided.
The Torquer must be able to operate as a generator. For a shunt
machine, the field current is undirectional, therefore, the machine changes
from motor to generator mode of operation naturally if the back emf is
higher than the source voltage. A series motor will not do so unless its
series field is reversed through mechanical switching. In view of the
need for Torquer PCU switching and sequencing in any case, field reversal
can be coupled to changes in PCU operational modes.
4.2 Torquer Specification
The Torquer used in the EMT breadboard demonstration is a series
wound DC motor built by the General Electric Company. The rating of the
machine is as follows:
27 shaft horsepower (P 7200 rpm ; 180 VDC, 130 amperes
Measured resistances (25 C)
Armature .0174 ohm
Commutating field .0075 ohm
Series field .0087 ohm
Calculated inductances (unsaturated)
Armature 0.3 mh
Commutating field 0.14 mh
Series field 2.58 mh
70
-------
Temperature rise and cooling (at rated conditions)! 90 C with 150 CFM
at 2" HpO forced ambient air
Mechanical over-speed capability: 10,000 rpm
The machine has a shunt tickler field rated at 10% of the series
field ampere-turns. This was built into the machine in the event there
was difficulty in establishing flux for voltage build-up during regenera-
tion. The field was not used since the Torquer PCU inherently provides
this function through its precharge and commutation process.
Weight - 153 pounds
o
Armature inertia - .031 Ib/ft/sec
Dimensions 9 3/4" diameter x 20 1/4" overall length
A characteristic map of the machine operating at rated DC voltage is
shown in Figure 4.1.
4.3 Speeder Specifications
The machine was purchased from available equipment and con-
sequently not optimized for the EMT service.
Speed (rpm) - 1,000 10,000
Output (kw) 1.0 10.0
Stator current (amps RMS) 52.0 42.0
Full load voltage (volts RMS) - 11.0 138.0
Phase -- 3 3
Input torque (Ib-ft)-input 10 10
No-load voltage ld.4 144.0
Frequency (Hz) - -- 66 660
Field excitation (rated) 3.3 amps at 14 volts
Overspeed capability 15,000 rpm
Weight 37 pounds
Size 9 3/4" diameter x 8 1/2"
overall length
71
-------
160
r 8000
140
- 7000 -
120
- 6000 -
100
u
z
UJ
u 100
to
CO
ID
o
C£
O
80
95
60
90
•40
85
20
80
SPEED, TORQUE, EFFICIENCY
CALCULATED FOR 180V CONSTANT
SUPPLY VOLTAGE
RATED CURRENT
- 2000 -
- 1000 -
200 300
LINE AMPS
FIGURE 4.1
DC PERFORMANCE OF THE TORQUER
72
-------
5.0 DYNAMOMETER TESTING OF THE EMT SYSTEM
The EMT breadboard equipment was tested on a dynamometer during the
spring of 1970. A series of tests were conducted to evaluate overall
system and individual component efficiencies and to establish the magni-
tude of power flow through each of the paths to and from the propeller
shaft. The data was hoped to be useful in; understanding the major sources
of system inefficiencies, characterizing the nominal and peak levels of
component operation and exploring the impact of vehicle weight, battery
size and engine power on system performance.
5.1 EMT Dynamometer Equipment
The EMT dynamometer equipment is indicated in the block diagram of
Figure 5.1. Figure 5.2 shows a picture of the test stand. The engine, a
1968 VW 1600 cc ICE equipped with intake port fuel injection, drives
through a planetary gear set adapted from a conventional automatic trans-
mission. The Speeder and Torquer are connected through the gear box as
indicated in the photograph of Figure 5.3. The dynamometer load consisted
of a Clayton Manufacturing Co. power absorber to simulate aerodynamic and
road drag and a Clayton Variable Inertia Flywheel assembly to represent
the inertia of the vehicle.
The equipment was operated from an adjacent control area with the
operator controlling the power train through the pedal controller and a
brake pedal actuating a disc brake on the propeller driving the dynamometer.
The operator used a driver's aid consisting of a strip chart recorder with
an actual velocity-time trace taken from the LA-4 driving pattern. A copy
of the trace is to be found in Figure 5.4
5.2 Instrumentation and Data Processing
A block diagram of the instrumentation points is shown in Figure 5.5.
Signals from the various transducers were conditioned (attenuated or
amplified and filtered) and then recorded on analog magnetic tape. The
data was processed off-line as follows. The analog tapes were converted
to digital tapes. These tapes were then used as data inputs to a complete
data processing, computation and print-out program.
73
-------
FUEL INJECTED
ENGINE
1600 CC
PLANETARY SYSTEM
ADOPTED FROM
AUTOMATIC
TRANSMISSION
3 PHASE
ALTERNATOR
(1-10 Kw)
OPERATOR OR
COMPUTER INPUT
POWER ABSORBER
AND INERTIA
FLYWHEEL
ASSEMBLY
"27" HP
DC SERIES
MOTOR
MOTOR AND GENERATOR
PO'.VER CONDITIONING
AND CONTROL
LEAD -ACID
BATTERIES
FIGURE 5.1
EMT DYNAMOMETER DEMONSTRATION
-------
en
FIGURE 5.2
EMT DYNAMOMETER TEST STAND
-------
—I
en
FIGURE 5.3
EMT ROTATING ELEMENTS
-------
FIGURE 5.4 (Cont'd)
0 TO 350 SECONDS
350 TO 700 SECONDS
FIGURE 5.4
SAMPLE LA-4 SPEED TIME CHARACTERISTICS
-------
I?
0
§
10
700 TO 1050 SECONDS
1050 TO 1400 SECONDS
-------
o
§
1400 TO 1685 SECONDS
FIGURE 5.4 (Cont'd)
-------
ENGINE
RPM
MANIFOLD
PRESSURE
TACHOMETER
I. C.
ENGINE
PLANETARY
GEAR
BOX
PR(
SH
—
^^^
DPELL
AFT R
1
ER
PM
DYNAMOMETER
LOAD
SPEEDER
FIELD
CURRENT
SPEEDER
oo
o
SPEEDER
RECTIFIED
VOLTAGE
TERMINAL
VOLTAGE
SPEEDER
RECTIFIED
CURRENT
I
TORQUER
SPEEDER
PCU
SPEEDER
PCU
CURRENT
TORQUER
PCU
CURRENT
BATTERY
VOLTAGE
TORQUER
TERMINAL CURRENT
PEDAL
POSITION
TORQUER
PCU
i
OPERATOR
'COMMAND
BATTERY
CURRENT
FIGURE 5.5
-------
5.3 Computation Sub-Routines
The high level of radiated electromagnetic noise from the equipment
coupled with the long instrumentation lines between the dynamometer stand
and the recording equipment required a sizable degree of noise filtering
prior to analog recording. The computation of electrical power requires
the multiplication of two quantities: current and voltage. However, after
filtering signals represent the average waveform values and do not contain
all the information of the true "chopped" waveforms associated with TRC.
The average electrical power, VT is:
V(t) Kt)dt
T
o
where T is a time period associated with the chopper frequency.
For all but very special conditions, it can be shown that:
Vl f 7T
and consequently it was necessary to perform off-line experiments to
establish a power coefficient, K-|, defined as
l ~ VF
where K-, can be experimentally correlated with Torquer current, mode of
operation (drive or regeneration) and vehicle speed. Waveform oscillo-
grams of the voltages and currents at the terminals of the equipment were
taken. Several of these are shown in Figures 5.6 through 5.9. The aver-
age power was then determined by the following process. The current and
voltage waveforms of the oscillograms were divided into segments, and the
instantaneous current, i , and voltage, v , measured. The average power,
VT was determined as:
m=m
m
81
-------
TORQUER
CURRENT
TORQUER
VOLTAGE
BATTERY
CURRENT
BATTERY
VOLTAGE
DRIVE MODE
VEHICLE SPEED 15 mph
IAV=191.8 AMP; F.F.=1.103
IAV=104.6 AMP; F.F.=1.72
HORIZ=0.5 MS/CM
1=200 AMP/CM
V=200 V/CM
HORIZ=0.5 MS/CM
1=150 AMP/CM
V=200 V/CM
FIGURE 5.6
82
-------
TORQUER
CURRENT
TORQUER
VOLTAGE
BATTERY
CURRENT
BATTERY
VOLTAGE
DRIVE MODE
VEHICLE SPEED 45 mph
IAV=146.4 AMP; F.F.=1.019
HORIZ=0.5 MS/CM
1=200 AMP/CM
V=200 V/CM
HORIZ-0.5 MS/CM
1=150 AMP/CM
V=200 V/CM
IAV=152.5 AMP; F.F.=1.03
FIGURE 5.7
83
-------
VEHICLE SPEED 30 mph
TORQUER
CURRENT
TORQUER
VOLTAGE
BATTERY
CURRENT
BATTERY
VOLTAGE
IAV=79.3 AMP; F.F.=1.193
IAV=25.3 AMP; F.F.=1.56
HORIZ»0.5 MS/CM
1=80 AMP/CM
V=200 V/CM
HORIZ=0.5 MS/CM
1=60 AMP/CM
V=200 V/CM
FIGURE 5.8
84
-------
TORQUER
CURRENT
TORQUER
VOLTAGE
BATTERY
CURRENT
BATTERY
VOLTAGE
REGENERATION MODE
VEHICLE SPEED 45 mph
IAV=126.3 AMP; F.F.=1.311
HORIZ=0.5 MS/CM
1=200 AMP/CM
V-200 V/CM
HORIZ=0.5 MS/CM
1=60 AMP/CM
V=200 V/CM
IAV=61.7 AMP; F.F.=1.61
FIGURE 5.9
85
-------
Since the waveforms are periodic, only one cycle needed to be analyzed.
Figures 5.10 and 5.11 show the power factor coefficients for the Torquer
using two different batteries.
The ohmic losses in the Torquer and battery are the product of the
root-mean-squared current and the resistance of the component. Again,
the degree of filtering precluded any meaningful digital multiplication
of current signals. It was therefore necessary to develop another con-
version factor, the form factor, F.F., from the off-line experiments. F.F.
is defined as:
PC. !RMS
r. r. — •——
I
F.F. can be correlated with the average current, mode of operation and
vehicle speed. The numerical computation of the form factor was performed
in a manner similar to that of the power factor coefficient. The form
factors of the Torquer in drive and regeneration for two different batteries
are shown in Figures 5.12 and 5.13. The battery form factors are shown in
Figures 5.14 and 5.15.
Off-line experiments were also conducted to relate the shaft torques
on the Speeder and Torquer to the average current and voltages at their
terminals (in the case of the Speeder, the three-phase rectifier) and
shaft speeds. The Torquer in drive and the engine were calibrated using
the power absorber portion of the Clayton equipment; the Speeder and the
Torquer in regeneration were examined using a Ward-Leonard system and a
separate, calibrated DC drive motor. The engine output torque was related
to engine manifold pressure and shaft speed. The various factors and cali-
bration data was programmed into various computation sub-routines. A
summary of these routines is indicated in Figure 5.16.
86
-------
2.2
00
—I
2.0
.8
.6
1.4
1.2
TORQUER POWER FACTOR, K
45 AH BATTERIES ) np|VF MOnF
30AHBATTERIESJDRIVEM°DE
15 MPH
45 MPH
30 MPH
0
100 150
AVERAGE TORQUER CURRENT, AMPS
200
250
FIGURE 5.10
-------
00
CD
1.0
0.8
0.6
0.4
0.2
15 MPH
45 MPH
60 MPH
30 MPH
TORQUER POWER FACTOR, K,
] REGENERATION MODE
45 AH BATTERIES
30 AH BATTERIES
I
50
100
150 200
AVERAGE TORQUER CURRENT, AMPS
250
300
FIGURE 5.11
-------
.5
00
1.4
.3
o
u
< i o
11 i .^
8
1.1
i.o
100 AMPS REGEN
130 AMPS REGEN (RATED)
150 AMPS REGEN
75 AMPS REGEN
200 AMPS REGEN
75 AMPS DRIVE
100 AMPS DRIVE
?50 AMPS DRIVE"""0 13° AMPS DRIVE (RATED)
10
20
30 40
VEHICLE SPEED MPH
50
60
70
80
FIGURE 5.12
TORQUER FORM FACTOR VS VEHICLE SPEED
45 AH, SLI BATTERIES
-------
1.5:
1.4
1.3
O
i—
u
CtL
O
1.2
1.1
1.0
D.
100 AMPS REGEN
130 AMPS REGEN
150 AMPS REGEN
200 AMPS REGEN
75 AMPS DRIVE
••a.
150 AMPS DRIVE
100 AMPS DRIVE
"T— D- 130 AMPS DRIVE
0
10
20
30 40
VEHICLE SPEED MPH
50
60
70
FIGURE 5.13
TORQUER FORM FACTOR VS VEHICLE SPEED
30 AH, SLI BATTERIES
80
-------
2.2
2.0
1.8
g 1.6
I—
u
£ 1.41-
1.2 -
1.0 -
45 AH BATTERIES
D D 30 AH BATTERIES
D
MPH
'D 30 MPH
60 MPH ^ 60 MPH
15 MPH
30 MPH
50
100 150
AVERAGE BATTERY CURRENT, AMPS
FIGURE 5.14
I3ATTHRY FORM FACTOR VS BATTERY CURRENT IN DRIVE
CJ1
o
-------
2.0
ro
.8
1.6
OS
o
o
1.2
45 AH BATTERIES
O D 30 AH BATTERIES
60 MPH
60 MPH
15 MPH
.0.
0
50 100
AVERAGE BATTERY CURRENT , AMPS
150
200
FIGURE 5.15
BATTERY FORM FACTOR VS BATTERY CURRENT IN REGENERATION
-------
FIGURE 5.16
COMPUTATION SUB-ROUTINES
ENGINE
MANIrOLU *
PRESSURE
TORQUE
MANIFOLD
FUNCTION
ENGINE
TORQUE
SPEEDER ARMATURE
CURRENT
SPEEDER FIELD
CURRENT
SPEEDER SPEED
V
X
SPEEDER
TORQUE
FUNCTION
>SPEEDER
INPUT
•rnnniir
TORQUER ARMATURE
CURRENT
TORQUER SPEED
V
•\
TORQUER
TORQUE
FUNCTION
TORQUER
.. -^ f \ in r~T
TORQUE
TORQUER ARMATURE
CURRENT
TORQUER SPEED
FORM FACTOR
FUNCTION
RMS. TORQUER CURRENT
•> (DRIVE AND REGENERA-
TION)
TORQUER ARMATURE
CURRENT
TORQUER SPEED —
TORQUER TERMINAL
POUER FUNCTION
TERMINAL POWER CO-
-> EFFICIENT
BATTERY CURRENT
VEHICLE SPEED
T v
BATTERY
FORM FACTOR
FUNCTION
>RMS BATTERY CURRENT
93
-------
5.4 Computer Print-Out and Data Presentalton
The digitized tapes were processed by computer, programmed to print-
out the following information at 5 second intervals:
Instantaneous values (over a 0.1 second period)
0 all shaft torques and speeds
• all machine, PCU and battery terminal voltages, currents and
power levels
• ohmic and the total of other losses (windage, magnetic, etc.)
in electrical machinery and the battery
• the integral value of all machinery, engine, PCU and battery
power up to that time (with appropriate provision for separation
into drive and regeneration)
Q integral values of ampere-hours into and out of the battery
0 vehicle speed and acceleration
e distance traveled to that time
e total time spent in regeneration and drive up to that time.
At the completion of each run, the computer printed out data (sampling
at 0.1 seconds) on the frequency of occurrence of:
• Torquer voltage and current (drive and regeneration)
e Torquer shaft power in drive and regeneration
• Torquer electrical power in drive and regeneration
9 Torquer PCU electrical power input during drive
« Torquer PCU electrical power output during regeneration
e Speeder shaft power input
e Speeder electrical power output (at full wave rectifier)
a Speeder PCU electrical power output
• Battery voltage and current
« Battery input and output power
94
-------
5.5 Data Logging
In addition to the recording of component performance, the following
were logged for each run:
• inertia loading (equivalent vehicle weight) and power absorber
torque at 50 mph equivalent road speed
• total fuel consumed
• battery used
• specific gravity of selected cells of the battery before and at
the end of the run
• battery temperature
5.6 EMT Dynamometer Performance on LA-4 Route
Table 5.1 presents a summary of 18 runs of the EMT system over the
LA-4 driving route. The major system parameters which were explored were:
• Vehicle weight and drag
- 2000, 2500, 3000, 3500 and 4000 pounds equivalent inertia
settings
- power absorber settings at an equivalent 50 mph road speed;
approximately 10 hp for the 2000 pound vehicle, 19 hp for the
4000 pound vehicle
• Lead-acid batteries
- 32-217 ampere-hour, 6 volt golf cart batteries weighing
approximately 2100 pounds
- 16-45 ampere-hour, 12 volt SLI batteries weighing approxi-
mately 500 pounds
- 16-30 ampere-hour, 12 volt power supply batteries weighing
approximately 320 pounds
• Engine power settings
- 12-16 output shaft horsepower at 1800 rpm
95
-------
PERFORMANCE
RUN IDENTIFICATION NUMBER
Vehicle Weight
Battery Type
Manifold Press. " Hg
Distance (D) - miles
EICE/D Kw-Hr/Mi
ETIR/D Kw-Hr/Mi
ETOR/D Kw-Hr/Mi
ETID/D Kw-Hr/Mi
ETOD/D Kw-Hr/Mi
Regeneration Time (sec)
Drive Time (sec)
% Regeneration Time
% Drive Time
EBO/D Kw-Hr/Mi
EBI/D Kw-Hr/Mi
AHI/D (calc) Amp-Hr/Mi
AHO/D (calc) Amp-Hr/Mi
EGI/D Kw-Hr/Mi
ESPO/D Kw-Hr/Mi
ETLR/D Kw-Hr/Mi
EBLI/D Kw-Hr/Mi
EBLO/D Kw-Hr/Mi
ETLD/D Kw-Hr/Mi
ESL/D Kw-Hr/Mi
Fuel Consumption miles/gal.
AAH (calc) (charge)
AAH (sp. gr.) (charge)
EROAD/D Kw-Hr/Mi
ESOTI/D Kw-Hr/Mi
ESOBI/D Kw-Hr/Mi
EGI/EICE %
ESOTI/EICE %
EROAD/EICE %
ETOR/D + ESOBI/D Kw-Hr/Mi
[EBI - (ETOR+ESOBI)]/EBI %
HC 56
HC 52
HC 49
HC 47 HC 55
HC 53
HC 50
2000
Tro j .
7.0
9.91
.565
.199
.091
.126
.065
1263
416
75.3
24.7
.091
.199
.859
.485
.229
.140
.108
.014
.008
.061
.080
17.4
(.374)
(.440)
.210
.035
.105
40.5
6.2
37.1
.196
0
2500
Troj .
8.0
9.74
.514
.178
.081
.202
.111
1096
574
66.6
33.4
.159
.168
.793
.979
.217
.123
.097
.008
.039
.091
.094
18.4
.186
.440
.235
.043
.080
42.3
8.4
45.7
.161
+4.2
2500
Troj.
9.0
9.60
.525
.182
.084
.227
.127
1073
606
63.9
36.1
.184
.154
.716
1.100
.217
.126
.098
.007
.047
.100
.091
18.1
.324
.321
.235
.043
.083
41.4
8.2
44.8
.167
-8.5
3000
Troj .
8.0
8.91
.529
.202
.090
.261
.153
1031
539
65.6
34.4
.212
.181
.809
1.330
.227
.128
.112
.008
.065
.108
.099
16.8
.521
.761
.264
.049
.079
42.9
9.3
51.8
.167
+7.7
3000
Troj .
7.0
9.78
.561
.210
.107
.234
.139
1136
544
67.6
32.4
.189
.190
.845
1.130
.224
.114
.103
.008
.051
.095
.090
17.0
.285
.537
.264
.045
.089
40.0
8.0
48.8
.196
-3.2
3500
Troj .
8.0
9.39
.546
.178
.072
.323
.193
1025
665
60.5
39.5
.266
.169
.790
1.750
.235
.154
.106
.008
.097
.130
.081
17.6
.960
1.340
.295
.057
.097
42.7
10.4
53.0
.169
0
3500
Troj .
7.0
9.42.
.579
.182-
.084
.329
.198
1034
656
61.1
38.9
.279
.182
.833
1.740
.261
.154
.098
.009
.100
.131
.107
16.5
.907
.618
.295
.050
.104
45.1
8.6
51.0
.186
-2.2
-------
TABLE 5.1
SUMMARY OF LA-4 TESTS
96
HC 32
HC 57
HC 58
HC 59
HC 33 HC 27
HC 29
HC 43
HC 37 HC 45 HC 62
3000
Troj.
8.5
9.25
.550
.126
.066
.287
.179
807
872
j48.0_ _
52.0
.224
.140
.650
1.433
.194
.135
.060
.008
.056
.108
.059
17.3
.783
.986'
.264
.063
.072
35.2
11.4
48.0
.138
+1.4
3000
G.C.
7.0
9.79
.514
.199
.098
.252
.138
1114
571
66.1
33.9
.206
.188
.901
1.110
.202
.123
.101
.006
.019
.114
.079
18.3
.209
(.051)
.264
.046
.077
39.3
9.0
51.4
.175
+6.9
3000
G.C.
9.0
9.75
.480
.172
.077
.289
.153
1041
639
61.9
38.1
.237
.164
.798
1.280
.188
.114
.095
.005
.023
.136
.074
18.4
.482
.693
.264
.052
.062
39.2
10.8
55.0
'.139
+15.4
4000
G.C.
8.0
8.34
.509
.204
.091
.391
.232
830
545
60.4
39.6
.337
.126
.881
1.895
.191
.116 ,
.113
.006
.043
.159
.075
18.9
1.014
.750
.330
.054
.062
37.6
10.6
64.8
.153
-21.4
2500
Willard
8.0
9.5§
.535
.124
•067
.209
.119
895
775
53.6
46.4
.156
.147
.681
.854
.180
.134
,047
.007
.011
.090
.046
17.9
.173
.350
.235
.053
.081
33.7
9.9
44.0
.148
-0.7
2500
Willard
9.0
9.55
.504
.116
.089
.203
.135
929
751
56.2
43.8
.175
.137
.632
.985
i, - *
; .076
: .,027
.005
.014
.090
.098
17.0
.353
.472
.235
.028
.038
34.5
5.5
46.0
.127
+7.4
4000
Willard
7.5
9.44
.569
.140
.066
.450
.287
811
864
48.5 |
51.5
.361
.164
.780
2.290
.187
.188
•074 ;
.008
.056
.163
.000
17.9
1.510
..705
.330
' .089
.099
32.9
15.6
58.0
.165
-0.6
3000
Willard
7.0
9.75
.567
.217
.096
.248
.135
1132
544
67.6
32.4
.199
.204
.894
1.082
.244
.147
.121
.005
.018
.113
.097
17.2
.108
.210
.264
.049
.098
42.9
8.6
46.5
.194
+3.4
3500
Willard
7.0
9.70
.686
.164
.085
.296
.196
953
718
57.0
43.0
.242
.184
.859
1.420
.222
.135
.079
.019
.090
.100
.087
17.1
.561
.780
.295
.054
.081
32.4 '
7.9
43.1
.166
+9.8
4000
Willard
6.5
8.05
.569
.212
.100
.346
.208
815
510
61.5
38.5
.292
.199
.900
1.720
.235
.137
.112
.006
.038
.138
.098
17.5
.820
.802
.330
.054
.083
41.3
9.5
58.0
.183
+8.0
3000
Willard
8.0
9.69
.545
.208
.111
.244
.154
1074
606
64.0
36.0
.212
.180
.816
1.270
.211
.068
.097
.004
.026
.090
.143
18.3
.450
.775
.264
.032
.036
38.8
5.9
48.5
.147
+ 18.3
-------
NOMENCLATURE FOR TABLE 5.1
EICE/D - energy from engine shaft per mile
ETIR/D - energy into Torquer shaft in regeneration per mile
ETOR/D - energy out of Torquer PCU (to battery terminals) in regeneration
per mile
ETID/D - energy into Torquer PCU in drive per mile
ETOD/D - energy out of Torquer shaft in drive per mile
EBO/D - energy out of battery terminals per mile
EBI/D - energy into battery terminals per mile
AHI/D - integral ampere-hours into battery terminals per mile
AHO/D - integral ampere-hours out of battery terminals per mile
EGI/D - energy into Speeder shaft per mile
ESPO/D - energy out of Speeder PCU per mile
ETLR/D - losses in Torquer system during regeneration per mile
EBLI/D - losses in battery during charge per mile
EBLO/D - losses in battery during discharge per mile
ETLD/D - losses in Torquer system during drive per mile
ESL/D - losses in Speeder system per mile
AAH(calc) - ampere-hour change of battery as measured terminals per mile
AAH(sp. gr.) - ampere-hour change of battery as measured by specific
gravity readings per mile
EROAD/D - road energy requirement per mile (computed during previous
computer simulations)
ESOTI/D - energy out of Speeder which flows directly to Torquer per mile
ESOBI/D - energy out of Speeder which flows directly into battery
97
-------
5.6.1 Distance and Time
The dispersion of distances the simulated vehicle traveled during the
tests is very small, less than 10% of the average. Variations can be
attributed to; the variability of various drivers in accurately following
the driver's aid, and small errors in the computing process.
Total trip time variations are very small, related to the detailed
time of starting and stopping the data recording. Two runs, columns 11
and 17 are much shorter; during these tests the vee-belts driving the
inertia wheels became dislodged and the tests were stopped short of the
full duration.
5.6.2 Regeneration and Battery Charging
The time spent in regeneration; i.e., the Torquer acting as a gener-
ator, is always greater than 60% of the total driving time. The battery
therefore receives energy more than 60% of the time or is being charged
almost twice as often as it is being discharged. Battery charging comes
from two sources, directly from the engine by way of the Speeder and from
the propeller shaft through the Torquer acting as a generator. In the
latter case, energy comes from two sources: the engine's torque contri-
bution and kinetic energy of the vehicle as it slows down. The contribu-
tion of the two machines' outputs to battery charging are close to one
another, thus accounting in each case for about 50% of the total battery
charging power.
The contribution of regenerative braking to battery charging is
difficult to determine—no instrumentation or computation technique was
available for that measurement. However, there is no apparent correlation
between vehicle weight and Torquer regenerative output as would be expected
if regenerative braking were a substantial portion of the Torquer's total
regenerative output. Indeed, qualitative observations as to the need for
the mechanical disc brake during LA-4 decelerations would indicate that as
98
-------
the vehicle weight increased, more mechanical braking was necessary,
probably negating the effect of the additional kinetic energy. The
Torquer system used during the tests had a constant, limited maximum
braking torque capability in it. A redesign would tailor this limited
capability to reflect vehicle size—higher limits for higher vehicles.
5.6.3 Drive and Battery Discharge
The EMT system spent approximately 35% of its time in the drive mode,
i.e., the Torquer acting as a motor. (An exception to this is the 2000
pound vehicle run of column 1. In this case, the engine output was too
high, and additional regeneration time was used to absorb the excess
energy.) Figure 5.17 shows the average Torquer electrical input and
mechanical output are approximately proportional to the vehicle weight.
Acceleration power requirements are proportional to vehicle weight; the
relatively constant engine output of the tests can only provide base
traction loads and in all cases the Torquer had to provide the peaking
power.
The energy output of the Speeder is constrained (for a fixed engine
output power) by the speed relationships of the engine, Speeder and pro-
peller shafts and the driving cycle. Speeder energy available to the
Torquer is therefore invariant with vehicle weight or battery type.
In the tests, the engine power settings were not increased enough
with vehicle weight and as a result the depth of battery discharge in-
creases with increasing vehicle weight. (In column 1, the excess engine
power shows up as a net charging of the battery during the test.)
5.6.4 Battery Operation
Figure 5.18 summarizes the effect of battery capacity and vehicle
weight on the performance of the battery. There is a strong difference
in the percent loss (or discharge efficiency) of the smallest (30 A-H)
batteries and the SLI and golf cart batteries. Since the battery energy
requirement for a given vehicle weight is relatively constant, it must
be concluded that the discharge efficiency is related to battery construc-
tion. To a first approximation, the plate area of the batteries should be
proportional to the capacity. Plate area establishes (to a first order)
99
-------
z
o
LU
O
NOTE:
TOTAL ENERGY FOR
LA- 4 + TIME THAT THE
DEVICE IS IN USE
UER PCU DRIVE INPUT-,'
TORQUER INPUT IN REGENERATION
FORQUER DRIVE OUTPUT
TORQUER PCU OUTPUT IN REGENERATION
2000
2500 3000 3500
VEHICLE WEIGHT - POUNDS
4000
4500
FIGURE 5.17
100
-------
40
8 030
w ae.
1 of O 20
I2O"-
10
2000
2500
BATTERY LOSSES:
A- 30 AH SLI
D - 45 AH SLI
O -217 AH GOLF CART
3000 3500
VEHICLE WEIGHT
4000
4500
FIGURE 5.18
EFFECT OF BATTERY CAPACITY ON BATTERY LOSSES
-------
the current density and polarization losses; losses are generally pro-
portional to current density. The construction forms of the 30 A-H and
45 A-H batteries are similar, both designed for SLI duty, and the poorer
performance of the 30 A-H batteries is expected on the basis of capacity.
The comparison with the golf cart batteries is more difficult. These
batteries are designed for deep discharge service and thus their thick-
ness to surface area is higher than the SLI batteries. It is probable
that the useful surface area of the golf cart batteries under high loads
is no more than that of the lower capacity, 45 A-H batteries and the
additional capacity is of no use in hybrid service. All the batteries
exhibit similar performance during charge. The charging power is low
(relative to discharge) and invariant with vehicle weight (as long as
the engine power remains approximately the same). The difference in
plate construction is not as important as in discharge because of the
lower currents and less dependence on electrolyte diffusion.
The calculated ampere-hour change of the batteries (ampere-hours in
minus ampere-hours out) varies in a similar manner with the charge state
changes as measured by specific gravity readings. These are, however, no
good correlation between them or their relative differences. It had been
hoped that the difference would be a good measurement of Faradiac efficiency,
that is, the amount of current charge which went into the parasitic gassing
reactions.
5.6.5 Energy Flow Paths and Component Efficiencies
A summary of average component and system efficiencies for a 3000
pound EMT vehicle is given in Table 5.2. The overall system efficiency
for any decreases in battery charge state brings this number even lower.
A certain amount of caution must be applied to this number in that the
road demand is based on earlier computer simulations of an EMT system
in which the road energy for LA-4 operation was computed using inertia,
141
drag and a LA-4 velocity profile. There is no way at present by which
the road load required on the dynamometer to perform the same profile
can be compared to these computer numbers.
The energy which flows from the engine shaft is distributed to the
road through various paths, each of which contain dissipative elements which
degrade the total energy flow. For example only a small percentage (
102
-------
TABLE 5. 2 AVERAGE EFFICIENCIES OF SUBSYSTEMS OF VARIOUS
3000 POUND EMT VEHICLES ON LA-4 MISSION
EROAD'
OVERALL EFFICIENCY - ^ EICE j 50^
GEAR BOX EFFICIENCY 75^
SPEEDER EFFICIENCY 70^
SPEEDER PCU 95^
TORQUER MACHINE EFFICIENCY (DRIVE) 67S
TORQUER PCU EFFICIENCY (DRIVE! 88^:
S TORQUER MACHINE EFFICIENCY (REGENERATION) 65^
TORQUER PCU EFFICIENCY (REGENERATION) 88ro
Power out
BATTERY UTILIZATION
^ -
Power in 4. (AH-V_.)
ID
-------
of the engine energy flows directly from the Speeder to the Torquer.
Since the combined Speeder and Speeder PCU operate at an average efficiency
of about 65%, the 40% of the engine energy available at the Speeder input
is degraded to 26% at the Speeder PCU output, leaving only 16% of the
engine output for the battery.
A major source of inefficiency in all engine paths is the gear box.
Off-line experiments indicate that the present gear box operates around
75% efficiency. The gearing was designed about an existing planetary gear
system and contains idler gears that would not be needed in a redesigned
system. The box is completely filled with heavyweight oil to insure proper
lubrication—a better design would use sling or mist lubrication.
5.7 Statistical Representation of EMT-Dynamometer Operation
Section 5.4 described the distributional information computed at the
end of each test. Figure 5.19 through 5.28 present histograms for the run
of column 5 of Table 5.1, representative of the data. There are various
ways of describing this statistical data; this section will treat several
of them.
5.7.1 Cumulative Energy Distributions and Through-Put Power Levels
Cumulative energy distributions of some of the histograms were computed
and are presented in Figures 5.29 through 5.33. The distributions were
calculated by multiplying the value of the mid-point value of each interval
of the histogram by its probability of occurrence and summing the result
with the sum of all the products starting with the lowest power level
interval. The summation was then divided by the sum of products over the
entire power range.
The distributions of battery discharge and charge energy exhibit
characteristics primarily dependent on the vehicle weight. As the vehicle
weight increases, the range of charge and discharge power increases, re-
lated to the increasing available deceleration energy and the increasing
acceleration effort required. The discharge power capability of the
battery is the ultimate limiting factor and the distributions indicate
that the peak power density (watt/pound) of the golf cart batteries is far
less than that of the SLI type. For the required peaking power of hybrid
operation, the excess golf cart battery capacity has little effect in pro-
viding additional power.
104
-------
50
40
O 30
u_
O
o
en
UJ
D
o
UJ
o;
10
BATTERY VOLTAGE
VEHICLE WT - 3000 LB CAR
BATTERY TYPE - 30 AH
1
I
100
120
140
160 180 200
BATTERY VOLTAGE - VOLTS
220
240
260
280
FIGURE 5.19
-------
o
en
50
40
5
t-
__i
f
2
u.
O
30
20
O
LLJ
ce:
10
BATTERY CURRENT
CHARGE
-20 0 20
VEHICLE WT - 3000 LB
BATTERY TYPE - 30 AH
60 100
BATTERY CURRENT-AMPS
DISCHARGE
FIGURE 5.20
220
260
-------
o
x
u
z
LU
5
O
LU
a:
50
40
30
>L 20
u
z
10
BATTERY POWER-CHARGE MODE
VEHICLE WT- 3000 LB CAR
70.4% OF TOTAL TIME IN CHARGE
10
15 20 25 30
BATTERY POWER-KILOWATTS
35
40
45
FIGURE 5.21
-------
o
00
BATTERY POWER-DISCHARGE MODE
VEHICLE WT - 3000 LB CAR
BATTERY TYPE-30 AH
29.6% OF TOTAL TIME IN DISCHARGE
20 25
BATTERY POWER - KILOWATTS
FIGURE 5.22
-------
25
20
I"
u.
O
I
£ '0
LLJ
ID
a
SPEEDER PCU OUTPUT POWER
VEHICLE WT - 3000 LB CAR
BATTERY TYPE - 30 AH
»«^^^^^^^^»
I 1 1 1 1 1
2.0
3.0 -4.0 5.0 6.0
SPEEDER PCU PO'-VER OUTPUT-KILOWATTS
7.0
8.0
9.0
"I3URE 5.23
-------
25
20 —
15
O
^
£
z
LLJ
a
LLJ
C£.
10
o
-250
TORQUE* C
.:C-.E WT - 3000 LB
TE-.Y TYPE - 30 AH
-180 -110
REGENERATION
-40
30 100 170
TORQUER CURRENT-AMPS
— DRIVE
FIGURE 5.24
240
310
380
450
-------
25
20
Q
z
uj 15
5
i—
LL.
O
10
LLJ
ID
O
LLJ
OS
U-
0
TORQUER PCU POWER-DRIVE
VEHICLE WT - 3000 LB CAR
BATTERY TYPE - 30 AH
32.9% OF TOTAL TIME IN DRIVE
15
20 25
TORQUER POWER-KILOWATTS
JO
40
45
FIGURE 5.25
-------
25
20
15
o
> 10
O
LLJ
a:
0
TORQUER OUTPUT POWER-DRIVE
VEHICLE WT- 3000 LB CAR
BATTERY TYPE-30 AH
32.9°/£ OF TOTAL TIME IN DRIVE
0
8
12 16 20
TORQUER POWER-KILOWATTS
24
28
32
36
FIGURE 5.26
-------
251
2 20
Z
LU
0
LU
7 15
LLJ
s
I—
u_
o
^ 10
J-
u
Z
LU
O
LU _
a: 5
TORQUER INPUT POWER-REGENERATION
VEHICLE WT - 3000 LB CAR
BATTERY TYPE - 30 AH
67.1% OF TOTAL TIME IN
REGENERATION
10
15 20 25
TORQUER POWER-KILLOWATTS
30
35
40
45
FIGURE 5.27
-------
50
TORQUER PCU POWER-REGENERATION
VEHICLE WT- 3000 LB CAR
BATTERY TYPE - 39 AH
O
40
67.1% OF TOTAL TIME IN
REGENERATION
Z
LU
O
Z
LU
£
30
20
u
Z
10
' u
8
I
12
16 20 24
TORQUER POWER-KILOWATTS
28
32
36
FIGURE 5.28
-------
BATTERY INPUT POWER
BATTERY OUTPUT POWER
BATTERY CUMULATIVE ENERGY DISTRIBUTION
3000 POUND EMT
45 AH, SLI BATTERIES
10
15
20
25
POWER LEVEL-Kw
30
35
40
50
FIGURE 5.29
-------
lOOr
80
Z
o
60
o
O£.
40
3 20
INPUT ELECTRICAL POWER
OUTPUT SHAFT POWER
CUMULATIVE TORQUER ENERGY DISTRIBUTIONS IN DRIVE
3000 POUND EMT
30 AH, SLI BATTERIES
10
15 20
POWER LEVEL-Kw
25
30
35
40
FIGURE 5.30
-------
100
80
60
I
z
o
CD
0£
O
O£
>
\—
<
3 20
40
ELECTRICAL OUTPUT
POWER
SHAFT INPUT POWER
CUMULATIVE TORQUER ENERGY DISTRIBUTIONS IN REGENERATION
3000 POUND EMT
30 AH, SLI BATTERIES
0
10
15 20
POWER LEVEL - Kw
25
30
35
40
FIGURE 5.31
-------
co
ELECTRICAL INPUT POWER.
SHAFT OUTPUT POWER
CUMULATIVE TORQUER ENERGY DISTRIBUTIONS !N DRIVE
4000 POUND EMT
45 AH, SLI BATTERIES
POWER LEVEL - Kw
FIGURE 5.32
-------
I
z
o
CO
g
SHAFT INPUT POWER
ELECTRICAL OUTPUT POWER
CUMULATIVE TORQUER ENERGY DISTRIBUTIONS IN REGENERATION
4000 POUND EMT
45 AH, SLI BATTERIES
POWER LEVEL-Kw
FIGURE 5.33
-------
The Torquer drive and regeneration distributions for several vehicle
weights and batteries follow similar patterns as the battery charge and
discharge distributions.
Table 5.3 summarizes the through-put power levels for the Torquer and
batteries for an aggregate of three vehicle weights. The number in the
columns represents the power level below which the indicated percent of the
total energy flow occurred. For example, 50% of the total energy into the
Torquer PCU of the 2000 pound vehicle flowed into the PCU at 8.5 kw or
less. While the 50% and 90% Torquer power levels are proportional to
vehicle weight (doubling for a two-fold weight increase), the Torquer
power levels during regeneration do not increase as rapidly. As suggested
previously, the limited regeneration torque command built into the operator
pedal control, and the substantial contribution of engine-propeller power
to Torquer regeneration are responsible.
The battery discharge levels are approximately equal to the Torquer
PCU input values as expected on the basis of the small contribution of
Speeder-to-Torquer power. The battery charge power levels however, bear
little resemblence to the Torquer PCU output in regeneration. Since the
Speeder and Torquer share the charging power, and the Speeder contribution
is relatively invariant with vehicle weight, there should be a relative
insensitivity of battery charging with vehicle weight.
5.8 Summary of EMT-Dynamometer Testing
The dynamometer testing of the breadboard EMT system revealed many
important operational and performance aspects which were not evident or
only qualitatively understood prior to the testing. The major conclusions
relative to each of these aspects is listed below.
« The EMT system is readily controllable and operator learning time
(at least in terms of the ability to follow the driver's aid) is
a matter of a few minutes.
• The total installed capacity of the lead-acid battery has little
impact on the efficiency or performance of the system. Rather,
attention should be paid to the development of lead-acid batteries
which have a high power density at low or modest capacity.
120
-------
TABLE 5.3 THROUGHPUT POWER LEVELS (KW) FOR THE TORQUER AND
BATTERIES AS A FUNCTION OF VEHICLE WEIGHT ON THE LA-4 ROUTE
Torquer PCU Torquer Shaft Torquer Shaft Torquer PCU Battery Battery
Vehicle Input-Drive Average Output-Drive Average Input-Regen. Average Output-Regen. Average Discharge Charge
Weight 50% 90% Power 50% 90% Power 50% 90% Power 50% 90% Power 50% 90% 50% 90%
_ Jew ky ig ; kw
2000 8.5 18.5 2.5 7.0 15.9 1.5 3.5 12.5 3.8 3.0 7.9 1.9 10.5 21.5 4.9 7.3
3000 14.5 32.5 4.1 10.8 25.6 3.0 4.8 16.5 4.1 2.8 9-6 2.1 13.8 30.3 6.5 11.5
4000 20.0 40.0 7.8 15.8 31.4 4.8 5.8 19.0 4.2 2.4 8.8 2.2 21.0 38.0 6.5 12.5
-------
• The charging energy to the battery is split almost equally between
the Speeder and the Torquer used as a generator. Improvements
in both of these systems are equally important in terms of charge
maintenance.
• The amount of energy delivered to the Torquer directly from the
Speeder is presently a small portion of the total engine output.
This is due to the large inefficiency of the planetary gearing
and the Speeder system. With improvements in these latter sub-
subsystems, the Torquer will be able to draw more energy directly
from the engine path, and less will be demanded from the battery.
• Improvement in the gear box will allow a greater contribution of
direct engine to wheel power thereby reducing the electrical
traction demands and providing more energy for battery charging
through Torquer generator action.
e To insure good controllability at low speeds in both drive and re-
generation, the present Torquer system must utilize variable
frequency operation. Methods for implementing the more desirable
constant frequency system (see 2.2.2) should be explored.
• Component heating did not present any particular problems during
the testing. The Speeder was the only device requiring a cooling
fan and this requirement might be relaxed with a more efficient
machine. The Torquer was not actively cooled but did run quite
close to its rated temperature. Improvements in that machine's
efficiency plus installation in the ram air environment of an
automobile suggest auxiliary cooling may not be needed. The major
power electronics of the breadboard equipment were cooled by
natural convection. In a vehicle installation, the closer packag-
ing might aggravate the cooling problem, but this is not expected
to be severe.
t Battery thermal behavior was better than expected in that the bat-
tery temperature rise (in still air) was about 40-60°F/hour of
operation for the smallest (30 A-H) batteries. With improved com-
ponents requiring less battery drain and in a ram air environment
battery cooling would not appear to be a major problem. Gassing
losses in the batteries was small; the 30 A-H cells required weekly
topping off based on a ten-to-twenty hours per week operating
schedule. Typical charging currents are less than the 1C rate.
• There is a strong impact of battery behavior on the design of the
electromechanical and electronic equipment. As the voltage stif-
fness of the battery is increased (lower internal impedance)
commutation requirements, Speeder and Torquer losses are reduced
and subsystem controllability is increased.
t Regenerative braking of the vehicle does not contribute substantially
to charge maintenance. However, the high current battery loading
associated with regenerative braking appears to provide a driving
force necessary to prevent serious acid depletion in the pores of
the plates.
122
-------
• The electrical braking capability of the present EMT system did
not adequately reflect the various vehicle weights explored. The
operator pedal controls should be modified so that the maximum
braking effort of the "accelerator pedal" is similar to that
braking of conventional cars, while the brake pedal should control
both the mechanical and electrical braking of the vehicle. In the
latter case, the braking effort should be raised nearer to the
maximum capability of the electronic/electromechanical components.
t The present Speeder is a relatively poor machine, with excessive
windage and magnetic losses. A reduction in its peak speed of
operation and a modification of motor construction and material
would reduce some portion of these losses. Reduction of maximum
speed will also reduce those losses internal to the gear box.
t The operation of the Torquer with a TRC controller introduces
additional losses in the machine. Redesign of the machine to
minimize these losses for EMT service is required.
123
-------
6.0 ADVANCED DESIGN STUDY OF TORQUER SYSTEM
6.1 Torquer Drive System Analysis
A simplified schematic of the Torquer system is shown in Figure 6.1
Both drive and regeneration mode use essentially the same components and
power control is based on the similar principles. Appendix A-2 presents
the details of the analysis leading to the programming of the circuit
equations on a digital computer. The analysis divides the total chopper
period, T, into several subdivisions. The dynamic circuit equations for
each interval are derived and integrated. The resulting equations for the
current flow during each time interval contains several coefficients whose
values are established through an iterative solution. Once a continuous
current solution is found, the program computes the following information:
t the time from period initiation for the commutating SCR to turn on
• the time from period initiation for the commutating and power SCR
to be turned off
t the time from period initiation for the commutating capacitor to
discharge
• peak and average current in the machine
• power output of the battery
• ohmic losses in the battery
• losses in the commutating circuit of the PCU
t theoretical power output of the Torquer (average current times
back emf)
• Torquer efficiency (theoretical power output/power output plus
machine ohmic losses)
t overall Torquer system efficiency (theoretical power output/power
output plus battery, PCU and machine ohmic losses)
• Fourier analysis of the Torquer current waveform
124
-------
BATTERY
MOTOR
ARMATURE
TORQUER SYSTEM IN DRIVE MODE
BATTERY
MOTOR
ASMATURE
FIGURE 6.1 TORQUER SYSTEM IN REGENERATION MODE
125
-------
6.1.1 Results of the Torquer Drive System Analysis
Typical computer print-outs of the analytical model are shown in
Figures 6.2 and 6.3. Figure 6.2 shows photographs of input parameter
displays and a pictorial display of the computed Torquer current wave-
form. Figure 6.3 contains a print-out of results computed from the wave-
form such as the maximum, rms and average Torquer current, losses, effi-
ciencies, and interval times. The lower picture of Figure 6.3 gives the
first few harmonics of a Fourier representation of the waveform, while
drawing in that picture is the current waveform constructed from the
Fourier components.
The analytical model represents the actual experimental waveforms of
the Torquer extremely well. For example, Figure 6.4 is an oscillogram of
the Torquer current and waveforms taken at conditions quite similar to the
actual waveforms shown in Figure 6.2 with the exception that the latter is
at a slightly higher chopper frequency.
The program input variables consisted of:
VB - battery open circuit voltage (200 volts for all cases)
E - Torquer back emf (assumed independent of the field current)
p - the equivalent resistance of the series motor including field,
m armature, brush drops
RD - resistance of the battery
D
L. - effective inductance of the series field
L~ - inductance of the linear inductor of the commutation circuit
C - capacitance of the commutating capacitor
R - equivalent resistance of the commutating circuit
T - chopper period
Hid X
t-, - the time at which the power SCR turns off
The computer results include the following:
8 - the time the commutation SCR is turned on, expressed as a percent
of the total period
126
-------
TORQUER MODEL INPUT PARAMETERS
PEAK CURRENT
259 AMPERES
AVERAGE
CURRENT
168 AMPERES
ZERO CURRENT
I CHOPPER
h«-PERIOD—I
I 1.0 KHz I
TORQUER CURRENT WAVEFORM SOLUTION
FIGURE 6.2
COMPUTER MODEL OF TORQUER SYSTEM
127
-------
T( :) = 1.16 -03
f.2?7-01
I(3)tfv ) : 2.^92+02
*K RMS ; : 1 . 6& +02
I( AVC J = 1 . 592+02
PtL«SS)= 3.175+03
P(0UT) = 1.592+OV
r/ = &. 337-01
P(b J = 1•A81+0f
P(C) = 7.02^+01
TORQUER SYSTEM PERFORMANCE
C M«.
COSINC
SINC
7.297-fOl
3.1 67-»-OC
I . 506-t-Ol
9.527+00
5. 3 SI+00
*.219+00
FOURIER COEFFICIENTS & RECONSTRUCTED CURRENT WAVEFORM
FIGURE 6.3
NUMERICAL PRINT-OUT OF TORQUER SYSTEM ANALYSIS
128
-------
VEHICLESPEED-30MPH
TORQUER SPEED - 1/2 RATED
AVERAGE CURRENT - 166 AMPERES
PEAK CURRENT - 280 AMPERES
CHOPPER PERIOD - 750 Hz
FIGURE 6.4
129
-------
I - maximum armature current
Inur - root mean squared armature current
RMS ^
I - average armature current
Form Factor - IRMS/Iave
P . . - total power loss in battery, Torquer and commutation circuit
P . - theoretical power output of motor, E I
out
"theoref PQUt + Pohffl1c
Table 6.2 presents data on the system for various parameter changes
at a Torquer back emf of 100 volts (approximately 1/2 rated speed or 30
mph). The parameter values selected represent those of the present equip-
ment and variations about the nominal levels. Columns 1 through 5 have
equal parameter values except for varying battery resistance. Columns fi
through 10 have varying inductance in the armature, while columns 11 through
14 examine the effect of changing duty cycle. In columns 15, 16 and 17
the chopper period is changed.
The effect of battery resistance variations is found in Figure 6.5.
The figure shows that variation in resistance has little effect on overall
system efficiency. However, this result is misleading since decreasing
resistance increases the average machine current and increases the power
output of the machine. For equal torque output, the duty cycle must"be
increased with increasing battery resistance and efficiency falls. Note
also the relative influence of battery resistance must be duty cycle
dependent. At very low duty cycle, the battery is connected to the machine
for only a small fraction of the time and the ohmic losses in the armature
predominate. At high duty cycles, the battery resistance is in the cir-
cuit more often and may outweigh armature resistance losses.
130
-------
TABLE 6.2
ANALYSIS OF TORQUER SYSTEM PARAMETER INFLUENCE
Run Number
VB( volts)
E (volts)
Rm(ohm)
RB(ohm)
LA(henry)
LC( henry)
Cc( farad)
Rc(ohm)
Wsec)
Tjtsec)
a (*)
8 (%)
lm*(an*)
!nns(amp)
IAve(amp)
Form Factor
Pohmic(kw>
Pout(kw)
_ Pout
P t+P0hmi
1
200
100
0.05
0.05
2xlO"3
2xlO"5
io-4
0.01
2xlO"3
IxlO"3
50
43
243
217.8
217.4
1.0
3.83
21.74
85
2
200
100
0.05
0.01
2xlO'3
2xlO"5
ID'4
0.01
2xlO"3
IxlO"3
50
43
287
261.6
261.2
1.0
3.89
26.12
87.1
3
200
100
0.05
0.15
2xlO"3
2xlO"5
io-4
0.01
2xlO"3
IxlO"3
50
43
202.4
180
179.6
1.0
4.59
17.96
29.6
4
200
100
0.05
0.10
2xlO'3
2xlO"5
ID'4
0.01
2xlO"3
IxlO"3
50
43
224.3
201.2
200.8
1.0
4.48
20.1
81.8
5
200
100
0.05
0.20
2xlO'3
2xlO"5
ID'4
0.01
2xlO"3
IxlO"3
50
43
184.6
162.4
162
1.0
4.59
16.2
77.91
6
200
100
0.05
0.10
io-4
2xlO"5
io-4
0.01
2xlO"3
lx!0"3
50
43
598.5
319.5
25 J. 6
1.26
13.16
25.36
65.8
7
200
100
0.05
0.10
5xlO"4
2xlO"5
io-4
0.01
2xlO"3
IxlO"3
50
43
259.2
168
159.2
1.06
3.18
15.92
83.4
8
200
100
0.05
0.10
IxlO"3
2xlO"5
io-4
0.01
2xlO'3
IxlO"3
50
43
227.1
179.1
177.1
1.01
3.59
17.71
83.20
9
200
100
0.05
0.10
2xlO'3
2xlO"5
ID'4
0.01
2xlO"3
IxlO"3
50
43
224.3
201.2
200.8
1.00
448
20.1
81.8
10
200
100
0.05
0.10
5xlO'3
2xlO"5
io-4
0.01
2xlO"3
IxlO'3
50
43
200.6
190.7
190.6
1.00
4.08
19.06
82.4
11
200
100
0.05
0.10
IxlO"3
2xlO"5
ID'4
0.01
IxlO"3
2xlO"4
20
5.95
135.1
108.8
107.3
1.01
1.39
10.73
88.5
12
200
100
0.05
0.10
IxlO'3
2xlO"5
ID'4
0.01
IxlO"3
4xlO"4
40
25.95
227.1
201.7
201.1
1.00
4.54
20.11
81.6
13
200
100
0.05
0.10
IxlO"3
2xlO"5
ID'4
0.01
IxlO"3
6xlO'4
50
45.95
419.6
399.3
399.2
1.00
19.16
39.92
67.56
1/1
200
100
0.05
0.10
IxlO"3
2xlO"5
io-4
0.01
IxlO'3
6xlO"4
80
65.95
588.7
576.8
576.8
1.0
45.63
57.68
5 5'. 82
15
200
100
0.05
0.10
5xlO'4
2xlO"5
io-4
0.01
IxlO"3
5xlO"4
50
35.95
315
266
265
1.0
8.17
26.51
76.4
' ^ 1
200
100
0.05
0.10
5xlO'4
2xlO'5
io-4
0.01
1.25xlO'3
6.25x'0'4
50
38.76
299
240
238
l.jl
5.52
23. E2
73. -5
1 — n — 1
200
100
0.05
0.10
5xlO"4
2xlO"5
ID'4
0.01
2.5xlO"3
1.25xlO"3
50
44.38
284
175
162
1.08
3.29
16.23
83.15
-------
3001-
35
100
200
to
o.
U>
ro
U
U
UJ
U
v:
i
~>
O.
t—
O
at.
LLJ
8
0.
Of.
LU
O
50
100
a:
O
0L
0L
30
25
20
15
10
VB = 200 VOLTS
E = 100 VOLTS
LA =2x 10"3 HENRY
T =2x 10~3 SEC (500 HERTZ)
max
a =50%
R =0.05 OHM
m
I
0.05 0.1 0.15
ARMATURE CIRCUIT RESISTANCE - OHMS
FIGURE 6.5
ARMATURE CURRENT & EFFICIENCY AS A FUNCTION OF
BATTERY RESISTANCE
0.2
0.25
-------
Figure 6.6 shows the effect of armature circuit inductance on theo-
retical power, efficiencies and form factor. At low inductance levels,
the current in the inductor is not strongly opposed by the L ^ of the
inductor and the peak current is limited only by the resistance of the
battery-machine circuit. A large amount of power is transferred but at
poor form factor and low efficiency. As the inductance increases, the
form factor and efficiency improve, but the peak and average current are
reduced. Further inductance increases beyond 10 henrys do not sub-
stantially affect the system's performance. As the current variations
decrease, power drawn from the battery is utilized more efficiently giving
rise to a higher average power and current in the machine. In contrast
to this, the increased current level increases ohmic losses with a slight
degradation of overall efficiency.
The above observations do not take into account hysteresis and eddy
losses. Table 6.3 presents the first three Fourier components of the
current waveforms associated with the various inductance levels. Examining
the magnitude of the sine and cosine vectors (cm's), demonstrates the rapid
decrease in h.armonic content with increasing inductance. In Section 7.2
it is suggested that the hysteresis and eddy current losses are related to
the flux density and frequency of operation. Using relations cited in
that section, one would predict that the eddy losses of the fundamental
using the 5 x 10"3 henry inductor would be (|zf)2 that of the 10"4 henry
inductor. Hysteresis loss reduction would also occur, although the amount
would be a function of the exponent of the hysteresis function of Section
7. 2.
Figure 6.7 shows the effect of increasing duty cycle on the system.
As expected, increased power SCR on-time increases the power output of the
system. However, as the current levels increase the system ohmic losses
also increase and system efficiency degrades.
Figure 6.8 shows the effect of chopper frequency on system behavior.
In the Torquer (as well as the Speeder) the electromagnetic torque is pro-
portional to the average armature current. For conditions of constant
average current, the form factor increases as the frequency is decreased,
and ohmic losses in the copper conductors increase. Hysteresis losses
associated with the flux variations in the magnetic paths also increase
133
-------
700 r
600
500
CO
LLJ
OC
LLJ
CO
J>
400
OH
U
£ 300
ID
i—
5
Q£
200
100
0L
35
100
80
vP
o^
I
z
LU ,/•>
U 60
co 40
LU
g
o
i—
20
10
EFFECT OF ARMATURE CIRCUIT INDUCTANCE
VB = 200 VOLTS
E = 100 VOLTS
9
RB = 0.15 OHMS
"SEC (1000 HERTZ)
THEORETICAL POWER
OUTPUT
MAXIMUM CURRENT
AVERAGE CURRENT
5x10 "* '.0 " 5x10
ARMATURE INDUCTANCE - HENRYS
FIGURE 6.6
-------
600
500
100
90
80
70
E 60
50
55 40
D
O 30
20
10
r- ' 400
a;
U
LU
Oi
300
I
h-
z
oe
Of.
3
200
100
EFFICIENCY
MAXIMUM CURRENT
E • I
•
B
RMS CURRENT
VB= 200 VOLTS
E = 100 VOLTS
LA=2x 10~3 HENRY
-1
Tmox ' 10 SEC
RM + RB = 0.15 OHMS
Ou
0.2
0.4 0.6
DUTY CYCLE-a
0.8
1.0
FIGURE 6.7
ARMATURE CURRENT & EFFICIENCY AS A FUNCTION OF
DUTY CYCLE
135
-------
400
100 r
90
80
u 70
LLJ
y
u_
u_
UJ
UJ
60
50
40
P 30
en
UJ
D
^ 20
10
LU
o
0L
300
uo
LU
C£
UJ
Q.
z
UJ
£ 200
u
UJ
C£
D
Oi
100
0
10
-3
MAXIMUM CURRENT
AVERAGE CURRENT
VB = 200 VOLTS
E = 100 VOLTS
LA =5xlO"4 HENRYS
R., + RB = 0.15 OHMS
M D
a =50%
2xlO"3
CHOPPER PERIOD - SECONDS
FIGURE 6.8
3x10
-3
136
-------
TABLE 6.3
FOURIER COMPONENTS OF ARMATURE CURRENTS FOR DIFFERENT
VALUES OF SERIES INDUCTANCES
^ d u c ta n ce
Coefficient" ^
a (average)
a] (cos wt)
b,(sin wt)
a2(cos 2wt)
b2(sin 2wt)
a^fcos 3wt)
b,(sin 3wt)
j
c, = /a] +bi<-
c2=/a22+b22
C.,= /a32+b3^
Cl/a0 (%)
c2/a0 (%)
c3/ao (%)
henry
253.6
-241
117
33.8
38.6
-17
-22
268
51.3
27.8
106%
20%
11%
LA=5xlO-4
henry
159.2
-73
-15.1
-3.2
9.5
-5.9
-5.9
74.5
10
8.35
46.7%
6.3%
5.25%
henry
177
-35.6
-9.9
-1.8
5.3
-2.5
-3.3
37.1
5.6
4.15
21%
3.16%
2.35%
LA=2xlO"3
henry
201
-17
-5.9
-1.2
2.8
-1
-1.6
18
3.05
1.88
9%
15%
0.93%
LA=5xlO"3
henry
191
-6.9
-2.4
-0.38
1.1
-0.46
-0.77
7.3
1.17
0.89
3.8%
0.61%
0.47%
a0 + ai cos wt + b] sin
cos 2wt + b,, .sin 2wt + a- cos wt
b_ sin 3wt t . . .
137
-------
with poorer form factor as the B-H magnetic loop covers a larger area.
The hysteresis loss per cycle may increase, but the number of cycles per
unit time decreases. The overall loss at any frequency also depends on
the magnetic material used in the machine, (The magnetic losses in the
PCU are related directly to frequency, since they undergo a constant
magnetic loop per commutation cycle.)
6.1.2 Summary of Torquer Drive System Analysis
The following conclusions can be drawn from the above results:
t Overall system efficiency is a strong function of battery
resistance. The importance of battery impedance as it
affects the duty cycle per Torquer power output should be
recognized.
« The filtering effect of the armature circuit inductance impacts
in two ways, affecting the resistive losses for a given power
output as well as the magnetic losses. For the present system
an overall inductance of 1 x 10~3 to 2 x 10~3 henrys is desired.
• Increased duty cycle at constant circuit parameter values in-
creases the power output but also increases system losses. The
relative gains (and therefore system efficiency) are functions
of battery and armature circuit impedance.
• Lowering the chopper frequency increases the form factor at con-
stant current resulting in higher ohmlc losses.
6.2 Torquer Design
6.2.1 Loss Mechanisms in the Torquer
There are several sources of losses in the Torquer: windage, bearing
losses, brush friction, brush drop and ohmic losses, hysteresis and eddy
current losses. Of these, the first three are primarily mechanical design
considerations, and little flexibility is allowed in them. The brush drop
loss is associated with the contact of brushes and commutator. The first four
losses are estimated by General Electric (and in some areas supported by
TRW tests) to be:
• windage, bearings, and brush friction - 400 total watts at 7200 rpm
• brush drop - 200 watts at 130 amperes DC, 7200 rpm
138
-------
If the machine is fed from a pure DC power source, the remaining
losses at 130 amperes, 7200 rpm are:
• ohmic losses
- armature, 320 watts
- commutation windings, 130 watts
- series field, 170 watts
• hysteresis and eddy current losses, 700 watts.
The total estimated losses at rated current and speed amount to 1.92
kw, giving an anticipated efficiency slightly over 90%. The reasons for
the much lower average machine efficiency are given below.
In the breadboard equipment, the Torquer does not see a DC source or
load but rather a chopped voltage due to the Torquer PCU. As previously
stated, this pulsating terminal voltage gives rise to a pulsating current
which increases both ohmic and magnetic losses. The former effect was
described in terms of the form factor in Section 5.3. Form factor losses
are aggravated by the high current levels required in the Torquer's opera-
tion. Figure 6.9 shows a distribution of Torquer drive current levels
taken at 5 second intervals for the first few minutes of one of the LA-4
runs. Differing vehicle speeds are associated with each point. The
Torquer current exceeds its rated level most of the time and increased
ohmic losses and an efficiency less than rated must be expected.
The current variations also produce strong flux variations in the
stator and armature structure. These flux variations tend to induce eddy
currents and produce a magnetic cycling in the iron paths. While the
armature is comprised of laminated, low loss iron typical of DC motor
design, the field structure (stator) is not. In classical machine design,
the machine frame or yoke and pole cores are solid pieces and only the
pole faces are laminated to reduce the losses due to the localized effect
of armature slot flux ripple.
139
-------
FIGURE 6.9
TORQUER EFFICIENCY ON THE LA-4 RUN
80
^
(J
Z
LU
0
LI 50
u_
LU
Qi
LU
D
O
o
20
. ' • -*
•— y ** * • >XK v* *
V • * * \
•/ « • • ^
./ • . :?.2MPH
' •
®3.5
_ • .2.6 • ®2.86 MP H
JL
\ ®1.63MPH ®2.54
r
i
/* ^
1 LU
o:
1 3
ILU
1-
s
1
3500 LB
DRIVE MODE, ANY SPEED
1 1 1 1
100
200 300 400
ARMATURE CURRENT, AMP
500
-------
6.3 Torquer Design Improvements
There are two approaches to reducing the losses in a DC motor operat-
ing from (into) a chopped voltage source (load). Both approaches may have
additional cost penalties associated with their use. This section will
discuss these possible design changes and describe some problems which may
occur with their adoption.
6.3.1 Addition of Armature Circuit Inductance
Additional armature circuit inductance could be added in the series
winding and/or in an external inductor. If the field winding were increased,
there would probably be an increased penalty of machine weight and cost. On
the other hand, use of an external inductor allows the total inductance to
increase without affecting the motor design. The addition of an external
inductor represents an additional costly item that takes up additional space
and produces losses and heat.
No matter what the cost or weight penalty, additional inductance has a
dramatic effect on Torquer system behavior. A test of the Torquer system in
regeneration was made with and without an external choke. The choke was a
duplicate of the energy storage inductor used in the breadboard Speeder PCU,
having DC inductances as follows:
_3
1.8 x 10 henrys at 50 amperes
_3
1.20 x 10 henrys at 130 amperes
It was used in this test as a swinging choke; that is, it was expected to
be saturated at the rated Torquer current. Tests on the Torquer alone under
chopper control indicated the machine's inductance was drastically reduced
from the DC design levels (see Section 4.2). The following inductance values
were determined at low speed (approximately 25-33% of rated machine speed)
for 500 hertz chopper operation:
0.75 x 10" henrys at 60 amperes average
0.35 x 10 henrys at 130 amperes average
_3
0.25 x 10 henrys at 200 amperes average
With the addition of the external inductor a dramatic change was made in
the Torquer system's performance.
141
-------
Table 6.4 summarizes the results of the test. The computations were
made by numerically integrating values from armature current oscillograms,
For an equal average current, the output torque of the machine with the
inductor is nearly double that without it, and the machine efficiency is
over twice as large. The inductor losses are included in the PCU effi-
ciency calculation and thus the PCU with the inductor shows a slightly
lower efficiency; still the overall effect is to raise Torquer system
regeneration efficiency from 39.6% to 80.9%. The results for the drive
mode are similar.
TABLE 6.4
EFFECT OF AN EXTERNAL CHOKE ON TORQUER SYSTEM PERFORMANCE
IN REGENERATION (2/3 rated speed)
Without Inductor With Inductor
Average Torquer Current 76 76
Average Torquer Voltage 84.7 94.5
Shaft Torque, ft-lbs 13.0 10.2
Shaft Power, kw 10.0 7.8
Torquer Output Power, kw 4.0 7.2
Machine Efficiency, % 40 92
Battery Input Power, kw 3.78 6.25
Torquer PCU Efficiency, % 95.1 86.1
Torquer System Efficiency, % 39.6 80.9
The impact of additional external inductance is lessened at higher
speeds and currents. In either case the chopper duty cycle is increased,
resulting in a better current waveform. The added inductor therefore
need not remain unsaturated at these higher levels; indeed saturation may
be desirable to automatically decrease the armature circuit time constant
whenever fast response is required at high output power levels. The use
of a small, highly efficient swinging choke may be quite adequate for
satisfactorily improving the present machine's performance.
142
-------
Another possible design effect of additional circuit inductance would
be the reduction in the size of the present machine. With a better form
factor, the machine losses would be reduced, and a higher torque output
could be achieved for the same average current. Since losses and torque
output determine the machine size, Improvement in these areas might result
in a smaller machine capable of the same power performance.
It is probable that the external choke would improve the machine's
commutation. (It might be noted in passing that the present Torquer has
never exhibited any detrimental sparking and the commutator seems in
excellent condition after more than one year in service.) Reduction in the
peak current levels would reduce the reactance voltage and the lessened
current swings in the interpoles should make these latter elements perform
more satisfactory.
6.3.2 Lamination of the Poles and Yoke of the Machine
A second method of effectively increasing the DC machine's inductance
(and thus efficiency) in pulsating service is to laminate the poles and
yoke of the machine with low hysteresis iron. Such an approach is more
like that used in AC commutator motor design where the stator must operate
under pulsating flux conditions.
The problem of operating a DC machine from a pulsating voltage source
has arisen in railroad service where the power source is rectified AC. In
the field of industrial motor control many times the power source is a
polyphase thryistor bridge. When conventional DC motors are operated from
such sources, results similar-to those in the EMT are observed: loss in
efficiency and heating. In some cases additional commutation problems can
arise due to the introduction of transformer emfs in the armature coils
and the inability of the interpoles to compensate for these induced volt-
ages.
The approach to the pulsation problem generally has been as follows:
t Redesign the motor yoke and series field pole cores to include low
loss iron laminations.
• Laminate the interpoles with low loss iron.
0 Minimize spurious eddy current paths through bolts, rivets, pole
attachments, etc.
143
-------
Some concern may be expressed that the increased inductance of the field
winding will increase the transformer emf in the armature and degrade the
commutation process (shrink the black band current range of sparkless
machine operation). This need not present a problem, however, if the
interpoles are properly designed to insure that the interpole flux is in
phase with the armature current being commutated. While it is possible that
a spark developed by transformer emf action may act as hot spot to dis-
charge the reactance voltage, it seems quite likely that a machine designed
to minimize reactance effects which are most severe at high speeds and
currents, will not spark because of transformer emf at lower speeds and
currents.
[5]
A recent work by Byrne, et al, reports the design and testing of a
modified 3 hp, 1830 rpm, 24 volt machine. The machine was rebuilt using
transformer grade steel sheet laminations for the yoke and additional
modifications were made to the armature and pole shapes. No interpoles
were used. Using data presented in the paper, the following conclusions
can be drawn:
• The new design increased the pole flux per ampere by 40% and in-
creased the rating of the machine from 3 Bhp to 5.6 Bhp.
• The specific power density (kw/lb) was raised by 136%.
• The machine's efficiency was increased at all torque and speed
levels above that of the conventional machine.
- 167% increase at half rated torque, 10% rated speed
- 123% increase at rated torque, 10% rated speed
• The machine's efficiency was made to be almost independent of
torque level indicating an almost complete absense of any wave-
form induced losses.
[e]
Another work reported by Kawai and Myake report the modification of a
traction motor for use on the Japanese Tokaido Railroad. The authors
report that lamination of the yoke, main pole and interpoles produced
dramatic reduction in the machine's temperature rise, the effect being
more pronounced as the harmonic content of the waveform increased. While
it appeared to be necessary to reintroduce some dampening (shunt resistor)
in the main winding to completely insure sparkless operation, it is not
clear that a modification of the machine interpoles or armature would not
144
-------
have served the same purpose without the parasitic loss in the main wind-
ing.
6.3.3 Advanced Design Conclusions
It has been demonstrated by TRW and reported in the literature that
major efficiency increases can be realized by incorporating modifications
of a DC machine design/circuit inductance. While it is not confirmed at
this time which approach (or combination of approaches) is the best from
the standpoint of highest efficiency per unit cost (at fixed torque and
power performance), the laminated motor approach is favored. Further,
information must be gathered from machinery vendors before a final design
can be selected.
6.4 Speed Charging of the Torquer
The present Torquer runs about 1/3 rated speed for LA-4 operation
(^20 mph average speed). Its average power level is nearly rated and thus
its currents generally exceed its design level. One means of increasing
efficiency is to reduce the average current demand and increase the chopper
duty cycle by allowing the machine to run closer to its design speed. It
is possible to incorporate a gear ratio shift between the Torquer and the
propeller shaft so that the Torquer runs at higher average speed during
Mode One and thus is near rated speed at mode transition. Assuming the
limit case of rated speed at mode transition (^40 mph), the Torquer would
then run on the average at 1/2 rated speed in Mode One. For the same power
delivery, the current would decrease from 300% of rated to 200% of rated
and neglecting any changes in form factor, the new ohmic losses would be
approximately 4/9 of the present losses.
The gear shifting could be incorporated with the automatic functions
of mode transition; however, even under these conditions there would be an
increased weight and cost penalty due to the added mechanical and control
elements. Also, the higher speed would result in higher back emfs, limit-
ing the ability to overload the machine for acceleration maneuvers. It is
probable that the advanced machine designs indicated in the previous section
will result in substantial Torquer system improvement. If the level of
performance can be raised using those techniques, the need for speed chang-
ing may disappear.
145
-------
6.5 Chopper Frequency Selection and Power Control Improvements
The analysis of the Torquer drive system (Section 6.1) demonstrated
the desirability of high (chopper) frequency (1 khz). There are several
qualitative advantages in that choice of frequency. Some of these are:
e Generally the size and weight of filter components having an
equivalent filtering effect vary inversely with operating
frequency. Higher frequency will afford easier packaging
and mechanical design.
• As the operating frequency increases, the chopper period
diminishes with respect to the electrical time constants
of the motor and the form factors and efficiency improve.
For equal power output schedules, the motor size, weight
and cost can be reduced.
On the other hand, there are advantages to the PCU design in using
a low frequency:
a The PCU efficiency is improved as the frequency is lowered.
The losses associated with the turn-on and turn-off of the
semiconductors is fixed and as the period increases, these
losses assume a lessened role relative to losses associated
with saturated state conduction.
• The times required to turn the power SCR on and off is fixed
by the device's characteristics. As the chopper period in-
creases, these times become relatively smaller and the dead-
band of controllability decreases. A long period is desir-
able in order to generate low (approaching zero) and high
(approaching 100%) duty cycles.
The advantages of high frequency operation outweigh those of low
frequency since the improvement of machine performance and potentially
lower cost equipment outweigh those of a higher efficiency and dead-band
elimination in an already efficient PCU, particularly if other control
techniques can be used.
The present Torquer PCU conducts some power to the Torquer through
the action of the commutation circuit even under minimum torque command.
The continuous operation of the existing commutation circuit has the
advantage of reliability since the commutation SCR gate is always trig-
gered and allows commutation power to be utilized in the motor. On the
other hand, the commutation energy delivered per cycle is fixed and as
the chopper frequency increases, the commutation power increases, increas-
ing the minimum torque which can be controlled.
146
-------
The desire for a high chopper frequency is therefore at odds with
the problem of commutation power control and in the present system, both
are compromised by shifting the chopper frequency as shown in Figure 6.10.
In assessing ways of circuit improvement it is worthwhile to examine
quantitatively the impact of the 1 kHz chopper frequency.
The commutation circuit provides power to the motor by delivering
energy stored in the commutation capacitor (see Appendix A-2). In addi-
tion to the current flow to the capacitor during recharge, there is a
path for current to flow from the battery to the motor via SCR~ during
the period of oscillation just preceding the recharge. However, calcula-
tions show the maximum level this current can obtain in the present
configuration is 13 amperes—a value incapable of producing usable motor
torque. There is no reason to add additional or modify the existing
circuitry to isolate the commutation circuit during the oscillation
period and prevent this current from flowing.
On the other hand, it can be shown that at very low speeds the aver-
age armature current due to the discharge of the commutating capacitor is
inversely proportional to the Torquer speed and directly proportional to
the PCD switching frequency. For example, with the present commutating
capacitor (80 yF), commutation power develops rated current in the Torquer
at 340 Hz at 5 mph and at 1 kHz at 15 mph. If the commutating capacitor
could be reduced, rated Torquer current due to commutation energy would
be produced at a higher frequency and lower speed. Thus, if the capacitor
could be reduced to 20 pF, rated current would be produced at 1 kHz and
5 mph, and a constant frequency system would be more practical. (The
lower capacity of the commutating capacitor would also reduce the size and
cost of that device.)
The principle obstacle to a lower capacity commutation capacitor is
that enough charge must be stored on it to turn the SCRs off. For example,
in dropping from 80 to 20 yF, the turn-off time of the SCRs must fall from
40 to 10 u seconds. Such high speed characteristics are presently avail-
able in SCRs as long as the -nr, -rp voltage levels and dissipation are
relatively low as they are in the EMT. In an upgraded Torquer PCU, these
faster semiconductors will be used.
147
-------
oo
CHOPPER FREQUENCY SCHEDULE OF BREADBOARD TORQJER PCU
— — — CHOPPER FREQUENCY SCHEDULE OF ADVANCED PCU
I
10
20 30 40
VEHICLE SPEED (MPH)
FIGURE 6.10
TORQUER PCU FREQUENCY SCHEDULES
50
60
70
-------
6.6 Torquer System Failure Analysis
The EMT dynamometer system was designed and constructed to prove the
basic EMT concept. Safety systems were provided to protect the personnel
and equipment within the context of a research and development environment.
Due to the continuous close monitoring and supervision of the equipment
during testing, automatic self-resetting features were not Incorporated so
that manual control of the many functions of the system was preserved in
contrast to the turn key operation essential for a vehicle.
The types of possible failures are grouped roughly into three basic
categories, dependent upon their final effect on the vehicle system. The
three types range from a malfunction almost undetected by the operator to
a failure which disables the vehicle entirely. The three types are as
follows:
• abnormal performance (automatic recovery)
• abnormal performance (degraded functional performance)
• catastrophic component failure (vehicle disabled)
Abnormal performance with automatic recovery is that type in which
there 1s no degree of disability relative to the driver or vehicle; rather
the operator will feel a momentary deviation (less than 1 second) from
the expected vehicle behavior. The third type of failure 1s the extreme
opposite in that the vehicle cannot operate on Its own and must be towed
to a garage. The degree of this disability 1s total. Between the two
extremes mentioned above there are a few types of abnormal performances
which only degrade vehicle functions, but still allow the vehicle to be
operated under limited conditions until service 1s reached. The degree
of disability can range from almost undetectable decrease 1n drive power
to no drive at all. For example, even with no power from the Torquer the
vehicle can still creep forward on the engine so that this 1s not a total
disability relative to total vehicle function.
6.6.1 Advanced Torquer Systems Incorporating Failure Protection
An advanced Torquer system 1s Indicated 1n block diagram form in
Figure 6.U. It reflects the advanced solid state switching system for
drive-regeneration configuration changes (see Section 8.1) and Incorporates
149
-------
FIGURE 6.11
ADVANCED TORQUER CIRCUITRY WITH SAFETY
PROTECTION & DETECTION
F1
2oov :
BATTERY
CROW BAR
*
THERMAL
ACTUATION
SCRr
CONFIGURATION!
; SWITCHING
SCRr
REVERSE
—i—o
TORQUER
ARMATURE
CR,,- -
FORWARD
DRIVER
CONFIGURATION
DEMAND
(LOGIC)
CURRENT
SENSOR
TORQUER
FIELD
TORQUE
DEMAND
CRr
Ml (RELOCATED)
/'
POWER SWITCH ING
.SCR,
OSCILLATION
FAILURE
._, DETECTOR *
I
COMMUTATION
CIRCUIT (SCR2)
J
{>
POWER SWITCHING LOGIC
CURRENT
REGULATOR
I
I
LOGIC PULSE |
WlpTH_MODULATpRj
REFERENCE
Ml *
SEQUENCER
IGNITION SWITCH
* ADDED CIRCUITRY
OVER CURRENT
DETECTOR *
-------
additional automatic safety and protection circuits. In the block diagram
power handling devices are separated into four basic groups:
• power switching (with commutation circuit)
• configuration switching
• Torquer
• safety devices
The operation of the power handling groups is controlled by three basic
groups of logic elements:
• power switching logic
- current regulator
- logic pulse switch modulator
• configuration switching logic
• Ml sequencing control (with thermal overload actuation)
While these groups of control logic do interact to some slight degree they
can be treated as independent blocks, i.e., a failure of a part of one
logic group only affects that group. For example, if the power switching
logic current regulator fails and demands current above the rating of the
PCU or motor, the Ml sequencing control is still operational and can per-
form its function to correct the overcurrent condition. The power handl-
ing elements are subjected to greater stresses, and their various modes of
malfunction must be considered on an individual basis.
The protection for a vehicle system must be effective both for the
vehicle as well as the driver. Uncontrolled acceleration or deceleration
cannot be tolerated, since either could place the vehicle and driver into
a condition where an accident can result. In addition to preventing
undesired vehicle behavior, the equipment cannot be allowed to destroy
itself by exceeding its power rating. This implies two basic methods of
protection, one to restore normal operations to the driver, if at all
possible, and another to prevent destructive power internal to the cir-
cuit.
151
-------
At this point, thought should be given to two methods of fault cor-
rection which have found a place in other applications. These are the use
of redundant circuits and manual overrides to bypass a failed component or
circuit. The use of redundancy is both costly and complex and should be
used only where the failure of a component would directly create a
potential for bodily injury. The dual braking system presently used in
automobiles is such an example. Redundance where the component failure
is merely an inconvenience is not anticipated in this type of vehicle.
Manually bypassing a failed device or overriding a safety mechanism until
help can be reached is somewhat impractical due to the complexity of the
system and the resultant large number of such bypass possibilities. How-
ever, it is possible with future on-the-road experience one or two such
overrides may be installed if a pattern of uncontrollable failures is
determined and the overall safety is not jeopardized.
6.6.2 Abnormal Performance - Automatic Recovery
6.6.2.1 Torquer PCU Power Switching SCRs Fail to Commutate Off
If the Torquer PCU SCRs fail to commutate while in regeneration, the
present circuit will affect the vehicle behavior in a manner dependent on
the speed of the vehicle. The specific speed at which one or the other of
two possible malfunction conditions can occur is a function of the wiring,
Torquer and semiconductor impedances as well as the actual burn-open cur-
rent of the fuse located in the Torquer circuit. All these components have
relatively large tolerances associated with them so that roughly below 30
mph one action takes place, and above 40 mph another.
Below 30 mph if the power switching pulse width modulator stops oper-
ating, a 100% duty cycle is commanded and maximum braking current will be
generated. At this speed the current will be too small to open the 400A
fuse in the Torquer armature. The vehicle will decelerate with maximum
braking torque. This condition cannot be corrected until the vehicle is
completely stopped at which time the drive configuration can be selected
and normal operation resumed. If drive is selected before the vehicle is
completely stopped, the configuration switching circuit will hang-up, the
consequence of which is described in the next section. Above 40 mph the
hang-up of the power switching SCRs will cause regeneration with current
sufficient to open the Torquer fuse, disabling the vehicle in both con-
figurations until a new fuse is inserted.
152
-------
The above discussion Indicates the need to locate Ml in the circuit
so that a simple failure to cormiutate in either drive or regeneration can
be corrected without the driver having to act. The proposed circuit shown
in Figure 6.11 provides this protection for all SCRs under both configura-
tions.
If the Torquer power switching SCRs (SCR1 and SCR.) fail to shut-off
to a misfire or temporary overload, a detection circuit will be provided
which will sense that the pulsewidth modulator has stopped operating.
This detection circuit will then open the main contactor Ml for a minimum
of 100 ms and close Ml again. The contactor will interrupt the current in
the SCRs and allow the semiconductors to turn-off. The effect on the driver
will be a momentary (less than .5 seconds) drop in power (in drive) or drop
in braking torque (in regeneration).
6.6.2.2 Torquer PCR Configuration Switching SCRs Fail to Commutate Off
If the Torquer PCR configuration switching SCRs (SCRp and SCRR) fail
to turn-off during the transition between drive and regeneration, or are
turned on out of sequence, there will be an uncontrolled surge of current
which will be detected by the overcurrent sensing circuit. This will
activate the fault clearing sequence of Ml, so that within 100 msec the
contactor will have opened, interrupting the current and closed again
reestablishing normal operation. If this occurs during configuration
switching the driver will only feel a slight (less than .5 seconds) delay
in transfer to the new configuration.
6.6.3 Abnormal Performance - Degraded System Functioning
6.6.3.1 Torquer PCU Current is Not Fully Controlled
If the current control (torque control) in the Torquer PCU fails slight-
ly off design conditions it is virtually Impossible to detect. When this occurs
the power handling devices will work properly, but the commands to the SCR
gate drive amplifiers will not perform properly. These problems can be
caused by the following:
• The accelerator pedal command does not call for the proper current.
• The current regulator does not regulate properly. The regulator
may provide for more or less current than that which the driver
desires. (This can be detected electronically by a redundant
153
-------
monitor, which can tell the operator and/or try to correct the
error itself. This circuit adds to the complexity and could
also malfunction, interfering with a properly working regu-
lator.) Although this is not a catastrophic failure, the
operator will probably observe a changing "feel" to the accel-
erator pedal and should be relied upon to call for service.
• If the logic pulsewidth modulator misinterprets the regulator
the effect and cure are the same as above.
If the current in the PCU reaches an uncontrolled level sufficient to
cause commutation failure, the procedures as described in 6.6.2 occur. A
sensor may have to be provided to warn the operator of the repeated cycl-
ing of the M] contactor after a prescribed number of openings in any given
period.
6.6.4 Catastrophic Logic Failure Leaving the Vehicle Completely or
Partially Disabled
6.6.4.1 Torquer PCU Does Not Switch Configuration
If the Torquer PCU fails to switch between the drive and regeneration
configuration one of three conditions is responsible:
• The configuration switching logic does not start to initiate the
sequence to transfer, so that the PCU stays in the previous con-
figuration but under full control. If this happens in drive the
vehicle will behave normally in drive and the vehicle can proceed
until service is found. The mechanical brakes will have to do
more work to control vehicle speed. If the PCU is in regenera-
tion, the vehicle can be operated up to a maximum speed corres-
ponding to a balance between road load torque and engine-propeller
shaft torque less the contribution of regenerative commutation
current in the' PCU.
« The configuration switching logic completes the first step in the
sequence to transfer, i.e., commands minimum power. This will
allow only commutation energy to flow. If this occurs in regenera-
tion or drive, the effect will be the same as in the previous
situation.
9 If the configuration switching logic sequences to the neutral
position and stops the PCU will handle no power at all. The
vehicle will be able to proceed similar to the previous situations
without commutation power.
154
-------
6.6.5 Catastrophic Component Failure Leaving the Vehicle Completely or
Partially Disabled
6.6.5.1 Torquer PCU SCRs, SCR, and SCR2> Short or Fail to Turn-Off Due
to Commutation Degradation
The failure of SCR, and/or SCRp to be turned off by the commutation
circuit, or the direct short of these SCRs will cause full current
in the system. This might result in full acceleration or deceleration.
One of two detection circuits will respond to open M].
Either the oscillation detector will indicate a failure, or the high current
will cause the sequencer to open M, for its prescribed length of time. M^
will close again having assumed the fault is cleared, but the failure will
repeat and an operator warning will be given. The above condition will
totally disable the Torquer system. However, the vehicle will still retain
its basic torque contribution from the ICE since the Speeder system will
not be affected and the vehicle can move slowly as in the last case of
6.6.4.1.
6.6.5.2 Torquer PCU Blocking Diode Short or Open
The Torquer PCU blocking diode, CRp, is one of the links from the PCU
to the traction battery. The path CRr-SCR, is the only one which does not
have the Torquer's impedance to limit the battery current when CRp develops
a short. Therefore, it cannot utilize a relay to interrupt the rapidly
rising currents which could develop. For this type of failure a fuse must
be used to halt the current before unreasonably high levels are reached and
thus F, is provided for this purpose. The fuse rating must be such that
faults in other semiconductors which can be cleared by M, should not
open the fuse and disable the PCU.
In a circuit with an inductive element such as the Torquer field wind-
ing, a path must be provided for the current flowing in the inductor to
dissipate itself in a load external to the inductor. Normally in the Tor-
quer system the inductive energy of the field passes through the flyback
diode, armature and/or battery. If the flyback diode, CRr, is open the
inductor will develop whatever voltage is required to preserve a path for
its current. This will result in an increase in the forward voltage across
SCR, until an avalanche breakdown occurs and current flows through CRR and
155
-------
CRD to complete the circuit. A large amount of commutation energy is still
available because the commutation capacitor will charge to the avalanche
voltage of SCR, and the system may go on operating in this manner until the
current activates the overcurrent sensor or causes a failure in the switch-
ing SCRs. In either case the M, sequencer will continue to cycle and warn
the operator.
6.6.5.3 Torquer PCU Configuration Switching Devices, SCRn, SCRR, CRn, CRR
Open Circuited u K u K
To change the configuration of the Torquer system from drive to re-
generation requires four semiconductors. The four devices are made up of
two diodes (CRQ and CRR) and two SCRs (SCRQ and SCRR). Since the Torquer
is a series machine, it requires regeneration current to flow before its
voltage can build-up. This inherent feature provides a degree of self
protection for failures if the semiconductor devices fail in the open
position. Since there will be no voltage generated, there is no need to
provide protection for driver, vehicle or PCU for this type of failure.
However, if the configuration switching semiconductors are not capable of
conducting current, the Torquer is completely disabled. If the fault
occurs during regeneration, there will be a lack of electrical braking
power and the driver must rely on his mechanical brakes.
6.6.5.4 Torquer PCU Configuration Switching Devices Shorted
The effect of developing a short in SCRp and SCRR are identical in
nature, so they can be treated as a group. In drive, current normally
flows through SCRQ, armature, CRp, and the field. If a short develops at zero
vehicle speed in SCRD, the path of current from the battery will be SCRn,
K U
SCRR (shorted), and field. The current flow will be limited by the
impedance of the field, the cable and terminal impedances, the semicon-
ductors and the battery impedance. Using values associated with the
present Torquer PCU, the current could rise to 760 amperes in 12 milli-
seconds. Similarly if SCRR is conducting during regeneration, and SCRD
develops a short the same conditions can occur as those caused by a
shorted SCRR. The manner of protection for both failures is to open M,.
156
-------
In Section 6.6.2.2 it was stated that if either SCRp or SCRR fail to
commutate off, then contactor M^ would clear the fault. The fuse FI must
be properly rated so that the current can flow through it until M^ has
opened if automatic recovery is to be achieved during a commutation failure
in the configuration switching circuit. The M-, protection circuit cannot
differentiate an actual short from a commutation failure. The differentia-
tion must be established through the repeated operation of the M^ sequencer
and operator warnings.
A short in CRR will produce extremely high current levels. The
armature will have almost no back EMF to oppose the potential of the
battery because the shorted CRR will bypass the current which normally
flows through the series field. Since the path for current when CRR shorts
has only F, for interruption, this device must be relied upon to clear
this fault.
In regeneration a short in CRp will deprive the field of current and
no armature voltage can be developed. The fault sensors cannot detect
this difficulty since none of the fault conditions have been reached. The
PCU will keep operating but no breaking torque will be developed. In drive,
a short in CRQ will cause no problems as it never has to assume a blocking
state.
6.6.5.5 Torquer Field or Armature Open
An open circuit in either the field or armature will not allow current
to flow from the battery in either portion of the motor or the PCUs while
in drive. The series connection of armature, field, configuration switch-
ing devices and power switching devices are inherently self-protected if
any of the elements develop an open circuit. Similarly in regeneration
there can be no destructive voltages or currents developed.
6.6.5.6 Torquer Field or Armature Shorted
Shorted turns in the power handling elements will produce an excessive
current and the development of hot spots. The amount of disability depends
on the percentage of turns shorted. Conceivably if only one is affected
the unit could continue to operate with only a small decrease in efficiency
which neither the sensors or the driver could detect. As more turns are
157
-------
shorted, internal heating will increase, and efficiency will decrease
until one or more events occurs:
e The driver detects the decreased power
• The need for higher power to produce the same mechanical work
activates the overcurrent circuit
• The motor thermal protection limit is reached by the internal
heating.
The current in the affected turns melts and the circuit opens.
6.6.6 Summary of Torquer System Failure Analysis
Table 6.5 presents a concise listing of possible Torquer system
failures, degree of disability and manner of protection. While this
list seems excessive it is to be noted that the majority of failures are
self-correcting or at least allow a small degree of vehicle mobility due
to the basic parallel configuration of the EMT.
No analysis of the impact of these electronic/electrical failures on
engine operation has been included. In the breadboard equipment the
engine throttle is tied to the Torquer current flow during Mode Two.
Failures which result in excessive current flows in either direction
would cause the throttle to open to its full position or completely close.
Complete loss of current would disable the throttle feedback system and
the engine would be left in its prefailure position. In any case, the
operator could use his brakes or ignition switch to bring the car under
control and back into the Mode One regime of operation. Since the engine
control system is still in its early development, further design effort
is needed to establish procedures and circuits which could be incorporated
into the system to prevent or override such failures.
158
-------
TABLE 6.5
TYPE OF FAILURES, PROTECTION & STATUS FOR THE
TORQUER PCU
COMPONENT OR
CIRCUIT
Torque r PCU
SCRl)
SCR2
SCR_ &
r\
SCRDJ
CRR
CRD
V
CRF
Commutation
Circuit
Driver Foot
Pedal Command
Configuration
Switching Logic
\ Current
Regulator
TYPE OF FAILURE
Failure to Commutate
Shorted or Degra-
dation of Commutation
Ability
Open
Shorted
Open
Shorted
Open
Shorted
Open
Does not provide
adequate reverse bias
to turn-off SCRl &
SCR2
Generates low cur-
rent command
Generates high
current command
Does not transfer
to drive
Does not transfer
to regeneration
Regulates at lover
than commanded
Regulates at Higher
than commanded
ICO
METHOD OF
PROTECTION
Automatic Reset
PCU will turn
off by Ml
None needed
Fuse Fl will
open, to turn
PCU off.
None needed
None needed
None needed
Fuse Fl will
open, to turn
PCU off.
Ml will open &
PCU turns off
PCU will turn
off by Ml
None needed
PCU will turn
off if current
exceeds over
current limit
by Ml.
None needed
None needed
None needed
PCU will turn
off if current
exceeds over
current limit
by Ml.
DEGREE OF
DISABILITY
None
Total
i
Total
Total
No
Regeneration
No
Regeneration
No Drive
Total
Total .
Total :
'
Drive & Re-
generation
Power Degraded
Total
No Drive
No
Regeneration .
Drive & Re- '
generation
Power Degraded
Total
,
1
1
-------
TABLE 6.5
(Cont'd)
COMPONENT OR
CIRCUIT
TYPE OF FAILURE
METHOD OF
PROTECTION
DEGREE OF
DISABILITY
Logic Pulse
Width Modulator
Range Limited to
Low Duty Cycle
Range Limited to
High Duty Cycle
Ml Contactor
& Logic
Torquer Motor
Armature
Does not open
Does not close
Open
Shorted Turns
Field
Open
Shorted Turns
None needed
PCU will turn
off if current
exceeds over
current limit
by Ml.
Crow Bar
Circuit will
turn PCU off.
None needed
None needed
None needed
None needed
None needed
Drive & Re-
generation
power degraded
Total
Total
Total
. Total i
Performance de-
graded. Ranging
from undetectable
to total dis-
ability
Total
Performance de-
graded. Ranging
from undetectable
to total dis-
ability
160
-------
7.0 ADVANCED DESIGN OF THE SPEEDER SYSTEM
7.1 Review of Speeder Performance on LA-4 Tests
The present Speeder system was shown to have an overall efficiency of
about 65% on the LA-4 route. The Speeder PCU has a relatively high effi-
ciency around 95%, while the Speeder itself is a poor performer. Figure
7.1 shows several points of machine operation at various speeds and power
levels taken during LA-4 operation. The machine's efficiency never exceeds
70%.
There are several contributors to the low machine efficiency; they are
• windage losses
t ohmic losses
t losses due to magnetic interactions.
Each of these contributors will be examined to assess their relative magni-
tude and to suggest means for their reduction.
7.2 Machine Losses
Figure 7.2 shows the losses in the Speeder due to windage and the
rotating magnetic field of the unloaded machine. Ignoring the latter,
the windage losses amount to 1.5 hp at 10,000 rpm and even at the speed
corresponding to the average vehicle speed (5000 rpm at 20 mph) these
losses are in excess of 0.5 hp.
The methods for reducing windage losses are well known. The machine
should be more aerodynamically designed to approximate a smooth rotor.
The present Lundell construction is particularly lossy and means should
be explored for filling in the interpole area. In addition, thought
should be given to running the machine at a lower top speed. Since
windage losses vary approximately with the cube of the rotational speed,
a small speed reduction would reduce aerodynamic losses substantially.
The lower speeds would also reduce gear box losses. On the other hand,
a reduction in speed would increase the torque (and current) rating of
the machine with some increase in machine weight.
161
-------
70%
£ 50%
IU
a.
20%
2,000 4,000 6,000
SPEED RPM
8,000 10,000
70%
UJ
y 50%
u.
UJ
Q
UJ
a,
20%
i
246
POWER INPUT Kw
10
FIGURE 7.1
SPEEDER SYSTEM EFFICIENCY
162
-------
FIGURE 7.2
1.6
1.5
1.4
CORE AND WINDAGE LOSSES
BOGUE SPEEDER
1.3
1.2
ex:
LLJ
£
o ]-°
Q.
LLJ
to
O 0.9
i
CO
to
o
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
o WINDAGE LOSSES
D CORE LOSSES AT 3 AMP FIELD
I
I
I
456
SPEEDER RPM X 10
7
-3
10
163
-------
The electrical and magnetic losses are more difficult to analyze.
When a polyphase synchronous generator is used to supply power to balanced
passive loads, both the generated EMFs and the phase currents are sinusoidal
and balanced. It can be shown that under these conditions the flux produced
by the armature current is in synchronism with the field structure (pole
faces and field windings) and no EMF or eddy currents can be induced in the
field. However, if the armature carries unbalanced or pulsating currents,
the flux produced by these currents will have components which no longer
rotate at the same speed as the field.
The stator current harmonics introduce alternating flux components in
the rotor structure which result in eddy and hysteresis losses in the rotor
pole surfaces and pole pieces. The flux components reflected back to the
stator can produce additional current harmonics in the stator windings,
creating additional ohmic and magnetic losses in the stator. It can be
shown that the hysteresis losses are proportional to:
fBmX; (0.5 < x < 2.3)
and the eddy-current losses are proportional to:
o t2 f2 B2
where f is the frequency seen by the field, B is the flux density, B is
the maximum flux density, a is the electrical conductivity of the field
structure and t is the thickness of lamination. Under normal conditions
of synchronous machine operation f = 0 (as far as the field structure is
concerned) and these losses do not occur.
In the EMT-Speeder system, the generator is operated into a rectifier
load which in turn delivers power to a switching load. The Speeder phase
currents and voltages therefore contain strong harmonics which cause large
losses in the machine, particularly in the field poles. The present Speeder
has a solid pole structure and is particularly susceptable to magnetic
losses. (The stator of course is laminated for alternating current service
and thus losses in the element are not increased as severely as in the
field.) Examination of the Speeder pole faces has shown discoloration of
164
-------
of the pole tips where flux crowding is expected. While the exact nature
of the discoloration was not determined to be merely the overheating of
the dip varnish or actual "bluing" of the metal, the losses apparently are
severe in this region.
7.2.1 Analysis of the Effect of the Rectifier on Speeder Performance
An analysis of the effect of the full-wave rectifier was made. In
the model, it was assumed that the rectifiers are ideal, i.e., turn on
and off instantaneously, and the load has infinite inductance so that the
phase currents appear as square waves. The conclusions of the analysis are:
• Due to rectification, the armature currents are non-sinusoidal.
As a result, an alternating current is induced in the field wind-
ing. The upper bound of this current is approximately 20% of the
rated field current (defined as the field current which produces
rated induced voltage at rated speed). The ohmic loss due to this
current is not severe. However, it may induce a fairly high volt-
age at the terminals of the field winding.
t If the rotor surface is not properly laminated, excessive ohmic
loss and local heating is produced. The frequency of the equiv-
alent damper winding current is six times the fundamental frequency
of the machine. (A damper winding is a short circuited winding on
the field poles of a synchronous machine. It is normally used in
larger machines as an aid in preserving synchronism by induction
action. In the analysis, the pole faces are assumed to act as
shorted circuits and thus can be replaced by an equivalent single
turn winding having the same effective resistance of the pole
faces.) At rated speed (10,000 rpm), the fundamental frequency is
about 600 hertz hence the damper winding current frequency is about
3600 hertz. At this frequency, the skin depth of the eddy currents
on the pole faces is less than one millimeter. The resistance of
the eddy current path thus becomes very high and in a very crude
calculation, the ohmic loss of the equivalent d-axis damper wind-
ing is estimated to be eight or nine times the field winding loss
at rated current.
• There are strong 5th and 7th harmonic components in both the intern-
al and terminal voltages on the stator side. The 5th harmonic is
approximately 30% and the 7th harmonic 65% of the fundamental com-
ponent.
• The electromechanic torque in the absence of any damping contains
pulsating components at six and twelve times the fundamental fre-
quency of the induced EMF. Its magnitude (the 6th harmonic) is
approximately 18% of the steady-state torque. The damping torque
of the present Speeder design is a function of the rotor pole
resistance. It is estimated to be less than 5% of its steady-state
component.
165
-------
e By introducing a low loss (high conductivity) damper winding on
the rotor side, the A-C component of the field current could be
substantially reduced resulting in a better waveform of the in-
duced voltage and a smoother electromechanical torque.
Typical phase current and line to line voltage waveforms for the
Speeder are shown in Figure 7.3. The Speeder and rectifier are operating
into a pure resistive load and the waveforms are symmetrical about the
zero axis. The waveforms were analyzed by the TRW on-line computer and
Fourier representations made. The analyses indicated the existence of
strong 5th and 7th harmonics in the current waveforms. The magnitude
of the 5th harmonic amounts to nearly 10% of the fundamental harmonic.
With the addition of the boost PCU, additional mechanism for Speeder
losses is introduced. The load on the rectifier is no longer passive but
contains reactive elements. When the Speeder power SCR is on, the storage
inductor charges and presents an inductive load. When the SCR opens and
the inductor current flows to the battery the load takes on a capacitivc
reactance character. The relative effects of these loads on current wave-
form and machine losses are further complicated by the necessity to vary
the PCU duty cycle with machine voltage, battery voltage and engine torque
variations.
Phase current and voltage waveforms for the Speeder-rectifier-PCU
system are shown in Figure 7.4. Speeds and average currents are the same
as in Figure 7.3. The lack of waveform symmetry about the zero axis is
apparent as well as the cycle to cycle variation of waveform. While a
Fourier analysis of these waveforms on a single cycle basis does not com-
pletely represent the total waveform, it serves to identify the major
harmonics. (There are undoubtedly subharmonics which would show up in
terms of a beat frequency between the machine frequency and the chopper
frequency.)
Table 7.1 presents the ratio of the various Fourier components to the
fundamental for the three waveforms of Figure 7.4. As in the case of the
pure resistive load, the 5th harmonic content is very high, approximately
10% of the fundamental. The effect of the rectifier appears to dominate
even with the addition of the PCU. However, the PCU does introduce
166
-------
FIGURE 7.3
50 AMPERES DC
93 VOLTS DC
5000 RPM
U
LU
o
30 AMPERES DC
97 VOLTS DC
5000 RPM
30 AMPERES DC
141 VOLTS DC
7500 RPM
ALL SCALES
30 AMPERES/cm
200VOLTS/cm
l.OMILLISECONDS/cm
167
-------
oe.
D
U
LU
2
£
50 AMPERES DC
5000 RPM
75 AMPERES/cm
200VOLTS/cm
0.5MILLISECONDS/cm
U
UJ
I
o
30 AMPERES DC
5000 RPM
30 AMPERES/cm
200 VO US/cm
0.5MILLISECONDS/cm
ae
Of.
D
U
UJ
<
30 AMPERES DC
7500 RPM
30 AMPERES/cm
200VOLTS/cm
0.5MILLISECONDS/cm
FIGURE 7.4
168
-------
TABLE 7.1
RATIO OF FOURIER COMPONENT TO FUNDAMENTAL (%) IN FIGURE 7.4
TOP FIGURE MIDDLE FIGURE BOTTOM FIGURE
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
llth
12th
cn
*ratio defined as — ;
ci
100
6.6
1.3
5.7
9.6
2.1
3.6
2.0
2.0
cl.O
2.2
2.7
where cn = ^
100
4.3
9.3
>1.0
11.7
1.0
10.6
5
2
1.5
3.0
1.3
i^ »:
100
4.5
6.6
2.3
12.5
2.0
9.8
1.2
2.3
1.9
2.2
1.3
169
-------
additional higher harmonics. These higher harmonics, although of lesser
magnitude than the 5th, occur at such high frequencies that their result-
ant eddy currents are restricted extremely thin layers on the pole face.
Up to this point hysteresis losses have not been considered. These
losses relate to the variations of the flux density, the frequency of
variation and the magnetic material itself. The pole magnetic flux is
always dominated by the main DC flux and the iron should be near or in
saturation. The harmonic flux variations in the pole faces are restricted
to a small layer due to the skin effect and their magnitude cannot be
expected to produce a large B-H area.
7.2.2 Summary of Machine Analysis
The Speeder operates in an environment where windage and magnetic
losses are extremely important. An advanced Speeder should be designed
along the following guidelines:
e lower speed range to reduce windage and gear box losses as well
as frequency of operation (the use of lower frequencies will
also allow the use of slower diodes in the full wave rectifier)
• higher air gap flux density to minimize the effects of armature
reaction
• introduction of low electrical conductivity pole faces by
- using laminated pole faces
- grooved pole faces on a solid pole
- anisotropic iron (having a grain orientation with a high
resistivity to eddy current flow).
7.3 Speeder PCU Improvements
The breadboard Speeder PCU control circuit was examined with respect
to possible improvements. A number of such improvements were identified
which would:
« employ control circuitry with linear characteristics
• permit the Speeder output to go to zero output
e utilize a constant frequency boost PCU
• minimize the current rating of the commutation SCR
170
-------
• minimize the commutation capacitor requirement
t employ the inherent short circuit safety of the alternator design
• delete the requirement for field control and thus provide true
constant field operation
t provide "back-up" overcurrent protection for the alternator above
and beyond that possible duty cycle control only.
7.3.1 Linear Control Circuitry
The present Speeder PCU uses a variable time pulser (VTP) control
logic. This logic, shown in Figure 7.5, utilizes a controlled current
source to charge a capacitor which is connected to the emitter of a uni-
junction transistor (UJT). When the voltage on the capacitor equals the
UJT trip voltage (Vn), the UJT transmits a pulse to the gate of the Speeder
power SCR. The power SCR then conducts until commanded off by a 1.0 kHz
sync pulse which also clears the timing capacitor. The power SCR off time
is thus coincident with the period during which the capacitor is charging
from zero to Vn and may be varied by adjusting the magnitude of the
charging current. The off time for this stage is given by:
t _ (0 (Vn) _ Kl
toff " I " I
where C and Vn are constants and I is the controlled variable. It is
apparent from this expression that the controlled period (tQff) does not
vary in a linear manner with the controlling parameter, I. Another cir-
cuit, the ramp and pedestal (RAP) control circuit, appears an attractive
linear system. A simple form of the circuit is shown in Figure 7.6. Tne
RAP circuit utilizes a constant current source to charge a capacitor which
is connected to the input of a voltage controlled trigger circuit. When
the voltage on the capacitor equals the trigger voltage (VprD) a pulse is
transmitted to the gate of the power SCR. The power SCR then conducts
until commanded "OFF" by a 1.0 kHz sync pulse which also clears the timing
capacitor. The power SCR "OFF" time is then coincident with the period
during which the capacitor is charging from zero to VpED and may be varied
by adjusting the magnitude of VpED. The off time for this stage is:
171
-------
FIGURE 7.5
VARIABLE TIME PULSER CONTROL
B+
O-
CONTROLLED
CURRENT
SOURCE
SYNC 1 kHz
PULSE
SCR GATE SIGNALS
UJT
FIXED
•O PEDESTAL
VOLTAGE
-fr-TIME
1 .0 ms
(TYPICAL)
4—>
'N,
172
-------
FIGURE 7.6
RAMP & PEDESTAL CONTROL
B+
VJ
CONSTANT SCR GATE SIGNALS fc
CURRENT r
SYNC
PULSE
1 KHz
••
1'
> TRinCFR O
cc
•-• ptl
vc
JTJUT
CONTROLLE
PEDESTAL
VOLTAGE, VpED
TRIGGER
VOLTAGE
PED,
PED,
— 1.0 ms
(TYPICAL)
7-1
TIME
173
-------
(C) (V )
= K V
I 2 VPED
where C and I are constants and Vpr-D is variable.
A comparison of the controlled time-off period (for constant frequency
operation) is shown in Figure 7.7. It is apparent that the RAP system pro-
vides a linear response to command inputs and is a preferred control cir-
cuit.
7.3.2 Zero Speeder Output
The present Speeder control system does not allow the Speeder output
to go completely to zero. This in turn means the Speeder cannot completely
unload the engine during cranking or engine control during operation. The
problem of the present design is due to the inability of the present system
to shut-off commutation when a zero output command is given. Examining
Figure 3.1 of Section 3.4, it can be seen that two components of energy are
drawn from the Speeder when the commutation SCR is pulsed "ON" once each
cycle whether or not the power SCR was in conduction during that cycle.
These two components are E.--the energy delivered through SCR to L, and
Ep--the energy delivered via the commutation capacitor. The components
can be derived as follows:
When SCRr is pulsed into conduction, current will flow in L equal to:
L*
v t
T _ s comm
ls ~ L
where t = n \jl C~~, the time needed for the commutation capacitor to
reverse its charge. Therefore
(V.) (n /TO
j _ O \* i*
*s " L
The energy stored in L due to this component is:
174
-------
off
0
roff
0
FIGURE 7.7
COMPARISON OF VTP & RAP CONTROL RESPONSE
VTP COMMAND CURVE
'off ~ K2 VPED
RAP COMMAND CURVE
COMMAND I
COMMAND V
PED
175
-------
? 7
Vs2 n2 LCCC
21
The commutation capacitor GC traverses a terminal voltage equal to 2Vg.
The energy consumed during this action is equal to
. cc N
EC - - 2
c = 2 f V
tC V B
The total energy per cycle (E ) taken from the alternator at minimum duty
cycle is equal to:
Es • EL + EC
2 ?
Vc n LrCr ?
F = s c c + ? r v 2
Es 2L Z Cc VB
The energy taken from the Speeder per cycle depends on V , which in turn
is a function of Speeder speed. The power delivered by the Speeder thus
depends on both V and the frequency of the chopper.
7.3.3 Constant Frequency Operation of the Speeder PCD
The previous section showed that commutation requirements limited the
ability of the Speeder output to go to zero. The present Speeder PCU
compensates for this limitation by adjusting to a lower frequency of
176
-------
operation if low outputs are needed. This variable frequency mode of
operation has the disadvantages outlined in Section 2.2.2.2. Advanced
Speeder PCU techniques should incorporate means for constant frequency
operation while retaining the capability of low power control.
7.3.4 Commutation SCR Rating
If the commutation SCR in the present system is pulsed "On" when the
power SCR is not conducting, a parasitic current, I , flows in the commuta-
tion SCR as described in 7.3.2. In turn the commutation SCR must be capable
of carrying an additional current above the normal commutation current,
and the higher rating means higher dissipation and cost.
An upgraded Speeder control circuit must include a logic gating net-
work which will always cause the power SCR to be gated on whenever the
commutating SCR receives a turn-on command. This technique will insure
that the power SCR conducts the load (inductor L) component of current,
while the commutating SCR will never carry more than its design commuta-
tion load.
7.3.5 Minimize the Commutation Capacitor Requirement
If the commutation capacitor does not have a sufficient charge when
the commutation SCR is pulsed on, the commutation will not succeed and
both the commutation SCR and the power SCR will remain in conduction. In
this case, both devices will have excessively large currents from the
Speeder. The present system has no protection against this mode of failure.
An upgraded PCU should include a voltage detector which monitors the
voltage on the commutation capacitor and compares it to a fixed reference.
Whenever the capacitor voltage is below a minimum required value, a lock-
out signal should insure that neither the power SCR nor the commutation SCR
can be turned on.
7.3.6 Utilize the Inherent Short Circuit Design Features of the Speeder
When the armature of an alternator draws current, a reaction component
of flux is induced in the field structure which tends to oppose the main
field flux. This effect, termed armature reaction, reduces the terminal
voltage. Under short circuit conditions, the armature reaction flux can
be so large in comparison to the main field flux that the net induced
177
-------
armature voltage is very small. The short circuit current is the induced
voltage divided by the winding impedance. The short circuit current of the
Speeder should be included as a major parameter in the selection and rating
of Speeder PCU components and in methods of fault protection.
7.3.7 Elimination of Field Control
The terminal voltage of the battery in the EMT system varies widely
due to changing load and charge state conditions. Voltages from 140 to 260
VDC have been observed. In order to work into this voltage range and pro-
vide Speeder torque and yet preserve the boost mode of PCU operation, the
present Speeder PCU incorporates a field control circuit which adjusts the
alternator field so as to maintain the Speeder output voltage below battery
voltage. This method has two shortcomings: (1) it requires additional
control circuitry, and (2) the capacity of the machine is not properly
utilized and thus it is larger than might otherwise be required.
The Speeder should be designed to have an output voltage at rated
speed and saturated field less than the minimum anticipated battery volt-
age. This redesign would allow the additional controls to be eliminated
and insure the machine is always used at full field. The disadvantages of
this approach is the more stringent current requirements on the remaining
control and power circuits as well as a slightly higher current rating of
the machine.
7.3.8 Overcurrent Control
The present control circuit employs a current sensing network plus a
field current regulator to limit the current drawn from the alternator.
This requires a complex circuit containing many components and is not
effective in the event a failure occurred in the control circuit.
An advanced system is needed containing a current sensing network
which operates a sequencer in series with the field. If the current rises
above a predetermined level, the circuit is opened for a fixed length of
time and then reset. This action will occur repeatedly for a predetermined
number of sample cycles, after which some fault indication should be pro-
vided to the vehicle operator.
178
-------
7.4 Preliminary Design of an Advanced Speeder PCU Control Circuit
A block diagram of an advanced Speeder PCU control circuit is shown
1n Figure 7.8. The essential elements are:
• ramp generator
• clock oscillator
0 pedestal comparator
• L-mult1pl1er
• error amplifier
• SCR gate drives
0 commutation lock-out.
The signal Inputs to the control circuit are:
0 engine tachometer speed
0 Speeder alternator current
0 commutation capacitor voltage.
The output signals are:
0 power SCR gate drive
0 commutation SCR gate drive.
The system employs a constant frequency ramp and pedestal comparator
in conjunction with a fixed reference and an error amplifier to control
the fraction of the ramp period during which the battery presents a load
to the Speeder. A network termed the "L-mult1pl1er" 1s Incorporated to
enhance system stability by preventing the Speeder current from changing
until commanded to do so by the tachometer signal. A protection network
denoted the "commutation lock-out" 1s provided to prevent either the power
SCR or the commutation SCR from conducting In the event an Insufficient
charge is stored in the commutation capacitor Immediately prior to SCR
conduction.
179
-------
PWR SCR GATE S1G
ALTERNATOR
CURRENT SIG
(DC INPUT)
JNGINESPEED^
^CONTROLLED5
PARAMETER)
ERROR
AMPLIFIER
MULTIPLIER
DC PEDESTAL SIG.
PEDESTAL
COMPARATOR
REFERENCE
DC LOCKOUT SIG
GO
O
REFERENCE
DC DRIVE SIG
PULSE DRIVE SIG
1.0 KHZ
RAMP
RAMP
GENERATOR
1 .0 KHZ PULSES
CLOCK
OSCILLATOR
AND INVERTER
COMMUTATION
LOCK OUT
COMM CAP VOLTAGE
(INPUT)
PWR SCR
GATE
DRIVE
COMMUTATION
SCR
GATE DRIVE
COMM SCR GATE SIG
DC LOCKOUT SIG
FIGURE 7.8
BLOCK DIAGRAM OF ADVANCED SPEEDER PCU
-------
7.4.1 Ramp Generator
The ramp generator circuit consists of a constant current source which
charges a capacitor at a fixed rate. The capacitor is discharged by means
of the clock oscillator at a rate much faster than the rate of capacitor
charging.
7.4.2 Clock Oscillator
The clock oscillator is composed of components which form a saturable
magnetic multivibrator whose operating frequency is adjusted to and held
constant at 500 Hz by the output of a positive voltage regulator. The out-
put of the clock is differentiated and rectified thereby producing a 1.0
kHz pulse train.
7.4.3 Pedestal Comparator
Duty cycle control is accomplished by applying the constant frequency
ramp signal to the sensing base of a trigger circuit while maintaining the
reference base at the pedestal voltage of the error amplifier. When the
ramp voltage is below the pedestal voltage, the comparator outputs a gate
signal and the power SCR is programmed "on". When the ramp rises above
the pedestal voltage, the comparator puts out a signal which turns on the
commutation SCR. The commutation circuit then turns the commutation and
power SCRs off. The SCRs remain off until the power SCR is again commanded
"on" by the next clock pulse.
7.4.4 L-Multiplier
The L-multiplier is a minor servo-loop which maintains the Speeder
current constant with respect to any EMT parameter changes which do not
cause the engine tachometer signal to change. This minor control loop is
employed since it can response much faster than the main error amplifier.
Because the minor loop attempts to hold the alternator current constant,
the energy storage inductance of the Speeder system appears to be much
larger than it actually is; hence, the name "L-multiplier". The output of
the L-multiplier determines the value of the pedestal voltage and is phased
such that changes in the Speeder current signal cause a change in the power
stage duty cycle so as to restore the reference current signal.
181
-------
7.4.5 Error Amplifier
The error amplifier monitors the tachometer signal and compares it
with a fixed reference. The output of the error amplifier is the reference
voltage for the L-multiplier. When the tachometer signal departs from the
desired value, the error amplifier transmits a corresponding voltage change
to the L-multiplier which in turn causes a shift in the pedestal voltage.
7.4.6 SCR Gate Drive Stages
The gate of the commutating SCR is turned on by signals from the
commutating SCR gate drive. The drive in turn receives its command signal
from the pedestal comparator. The drive stage is also connected to the
power SCR gate drive in such a manner as to pulse the power SCR "on" when-
ever the commutation SCR receives an "on" command. The power SCR gate
drive receives its initial command pulse from the clock by way of the
pedestal comparator. The drive puts out a steady "on" gate signal until
the ramp voltage crosses the pedestal voltage again when the gate signal
is removed.
To reduce the effect of commutation on the lower limits of power con-
trol a pulse skipping feature is used. When the pedestal voltage goes to
zero, commanding a zero power output, no gate signal is applied to the
power SCR. Since the previous function of the pedestal comparator was to
turn off the commutation SCR, both SCRs are turned off and remain in that
state until a finite pedestal voltage is applied.
7.4.7 Commutation Lock-Out
The voltage of the commutation capacitor is monitored and compared to
a fixed reference. If the capacitor voltage is below the required minimum
value, the pedestal voltage is forced to zero and holds the commutation SCR
gate pulse circuit inoperative.
7.5 Speeder System Failure Analysis and Protection
7.5.1 Speeder System Operation and Protection
The basic requirements of the Speeder PCU is to regulate the speed of
the engine by providing a variable torque load on it. The design of the
PCU is such that it cannot operate unless the traction battery is connected
to its output. With regard to the Speeder system, the battery is not a
182
-------
source of power as 1t 1s the Torquer, but 1s the output load which must be
present to absorb the Inductive energy stored 1n the PCU.
A block diagram of the Speeder system 1s shown 1n Figure 7.9 which in-
cludes the elements which are required to control the Speeder system in the
event of failure within that group. The power handling elements, alternator,
PCU, batteries and logic directly related to the PCU operation are essentially
the same as those used in the breadboard equipment. The fault protection
circuit allows power to flow from the alternator only when all safety condi-
tions are satisfied. In the protection system the power from the Speeder
1s controlled entirely by the presence or absence of the field current. The
conditions to be satisfied for field current to flow are:
t the battery voltage must be above a preset level
• SCR.J and SCRC must not be permanently on
t the Speeder current must exceed a preset level
• the engine ignition switch must be on
• the PCU or Speeder cannot be overheated.
The field sequencer portion of the fault protection circuit monitors
the above conditions and decides whether or not field current should flow.
(This element also 1s used to defeat the Speeder field during engine crank
1ng.) It functions by momentarily removing the possibility for electric
power generation and thereby allows the current flow in the SCRs to decay
and recover. The sequencer also decides when, after a series of unsuccess-
ful attempts to establish normal operation, to signal the operator that
the system may be falling and should be shut down.
The only other source of energy Into the Speeder PCU 1s that which
could flow from the battery. Relays are of no use 1n this application
since a reverse current flow Indicates a serious Speeder system failure.
Fuses are /therefore preferred. Since the protection system operates by
field control 1t is Important that 1t be tied to the engine throttle contro
Any extended field removal must also cause the throttle to cut-back or un-
acceptable engine (and Speeder) speeds may occur.
183
-------
FIGURE 7.9
ADVANCED SPEEDER CIRCUITRY WITH FAILURE DETECTOR
& SAFETY PROTECTION
ENGINE RPM-
CRF
^H-L
I
COMMUTATION
CIRCUIT (SCR2)
, 1_^psK5UILIBRIUMf'Ns^^
^ frjX^DETECTOR \S
•TXpKiriiMC
ENGINE
SPEED
ERROR
DETECTOR
•FAULT PROTECTION CIRCUIT
I
I
I
I
I
LOGIC
PULSE
WIDTH
MODULATOR
OSCILLATION FAILURE
I DETECTOR
OVERCURRENT
THERMAL
PROTEC-
TION OF
PCU
BATTERY VOLTAGE
DETECTOR
FIELD SEQUENCER
MANUAL
ON
f2
TO
LOGIC X
MODULES
• BATTERY
LOGIC
VOLTAGE
-IGNITION
-------
7.5.2 Abnormal Performance with Automatic System Recovery
7.5.2.1 Speeder PCU SCRs Fail to Commutate Off
If the Speeder PCU SCRs fail to commutate off due to misfire, temporary
overload, or temporary battery voltage collapse, a detection circuit senses
that the pulse width modulator has stopped oscillating. When this is
detected the field and logic current are disconnected for a minimum of 200
milliseconds and then reapplied. During this time the current in the SCRs
will go to zero and^as there are no gate pulses, they will turn-off. The
external effect will be a momentary increase in engine speed followed by
normal operation in about 0.5 seconds.
7.5.3 Abnormal Performance with Degraded System Functioning
7.5.3.1 Speeder PCU Does Not Control the Engine Speed Properly (within 25%
of reference speed)
The Speeder PCU might not be able to regulate the engine speed for the
following reasons:
• the logic amplifiers do not operate correctly
0 the PCU does not transmit the Speeder power
• the Speeder does not produce torque
t the engine torque increases or decreases to a level the Speeder
system cannot control. This failure would be associated with the
engine control or the engine itself and will not be considered here.
If the speed error detector, equilibrium detector, or logic pulse-
width modulator fail to perform their required function the regulated speed
of the engine can be affected. If the PCU attempts to regulate the engine
at a higher speed, the PCU would not recognize the engine's departure from
its reference speed. For example, a high engine speed implies a low current
in the Speeder and there is no protection for this condition. If the Speeder
forces the engine speed down the PCU would have to operate at a higher
current. In order to prevent a continued overcurrent condition, a thermal
interlock would remove the field current and the engine speed would in-
crease.
185
-------
7.5.3.2 Engine Controller Fails in Mode One
If the engine controller allows the engine power to rise, the Speeder
PCU will increase the Speeder armature current. Even if the Speeder does
not close torque due to armature reaction, a prolonged heating of the
Speeder or Speeder PCU will cause the thermal protection system to dump the
field current.
In the event that the engine power falls the PCU will decrease its
load as far as its control range allows. This of course may keep the engine
under control, but sustained low engine power will ultimately cause the
battery to deplete its charge.
7.5.4 Abnormal Performance with Complete System Stoppage
If the Speeder is unable to produce a reaction torque on the engine,
engine torque cannot be transmitted to the rear wheels and the engine must
be throttled down or another means for reaction torque must be provided.
If the engine is throttled down or shut off the vehicle can still proceed
for a short distance using the battery and Torquer if the engine can be
uncoupled from the drive train.
7.5.4.1 Shorted or Opened Windings
If the Speeder develops an interwinding short it will dissipate some
of the power which normally passes through the PCU. The speed regulator
will continue to adjust the duty cycle to keep the engine speed constant,
and there will be no apparent effect on the system operation until the
alternator overheats and is turned off and/or burns open the effected wind-
ing. Similarly if a winding opens up due to vibration or abrasion, the
Speeder will lose torque and the engine must be shut down.
7.5.4.2 Shorting of Speeder PCU SCRs
A direct short of the Speeder SCRs (SCR-j and SCR2) will produce the
same immediate effect as a hang-up, described in Section 7.5.2.1 except
that the reapplication of power will not succeed in establishing normal
operation. Rather the logic sequencer will try again to apply power, and
after a series of attempts the timing network will permanently turn-off the
alternator power. A warning signal will have to be presented to the
operator.
186
-------
7.5.4.3 Speeder PCU Block Diode Open
The Speeder PCU blocking diode (CRp) is the only direct link with the
battery. If this link is opened the energy in storage inductor will produce
whatever voltage is required to provide a path for its current. With CRp
open the only possibility is to force SCR, or SCR? to breakdown and conduct
in the forward direction.
7.5.4.4 Speeder PCU Blocking Diode Shorted
A short circuit in CRp will allow the battery current to flow backward
through CRp (shorted), and SCR,. The fuse between the PCU and the battery,
F,, will open and the alternator power will be removed.
7.5.5 Summary of Speeder Failure Analysis
Table 7.2 summarizes the above failure modes and contemplated courses
of action. Almost all Speeder failures result in the possibility of engine
overspeed and it is obvious that some means of overspeed protection must be
provided. In no case would a Speeder system failure present a real hazard
to the operator or vehicle.
187
-------
TABLE 7.2
SPEEDER PCD FAILURE DETECTION & PROTECTION SYSTEM
COMPONENT OR
CIRCUIT
Speeder PC'J
SCIU
SCR2
:RF
Commutation
Circuit
LI
Rectifier
Bridge
Current
Sense
TYPE OF FAILURE
Failure to Coramutate
Shorted or Degra-
dation of Commutation
Ability
Open
Short
Open
Does not provide
adequate reverse
bias to turn off
SCR1 & SCR2
Open
Shorted turns
Open
Short
Indicates high
current
Indicates low
current
METHOD OF
PROTECTION
Automatic Reset
Field Sequencer
will turn Power
off
None needed
Fuse Fl will
Open, to turn
PCU off.
Field Sequencer
will turn Power
off
None needed
Overcurrent
sense will
eventually
actuate field
sequencer.
None needed
Field Sequencer
will turn power
off
None needed
Thermal protec-
tion will turn
power off
DECREE OF
DISABILITY
None
Engine will
Speed up
Peak, current in
the PCU will
rise, possibly
beyond the com-
mutation cap-
ability.
Engine will
Speed up
Engine may drop
in speed until
thermal protectior
activates, then
;ngine will speed
up
188
-------
TABLE 7.2
(Cont'd)
DOMPONENT OR
CIRCUIT
TYPE OF FAILURE
METHOD OF
PROTECTION
DECREE OK
DISABILITY
ICE Speed Error
Uetsctor,
Current Regula-
tor
Pulse Width
Modulator
Commands less then
required current
None in PCU
(Engine should
be self-protect
ed)
Engine speed
will increase,
depending on
the degree of
failure
Commands more then
required current
Speeder Alternator
Armature ,
Field
Open
Shorted turns
Field sequencer
will turn power
off if over-
current is
reached
None needed
Engine speed
will decrease
depending on the
degree of fail-
ure
Engine will
speed up
189
-------
8.0 ADVANCED EMT POWER TRAIN OPERATIONAL CONCEPTS
This section describes work associated with three areas of advanced
system design:
• Solid state switching of the Torquer PCU between the drive and
regeneration configurations
o EMT power train start-up concepts
• System override approaches
8.1 Solid State Switching System
8.1.1 Three-Power Relay System
The original design of the Torquer PCU used three power relays to
switch the semiconductors and motor elements between drive and regenera-
tion. The power elements of that circuit are shown in Figures 8.1 and
8.2. The three relays are termed NL, M. and M,.. The relay M., is used
to establish the configuration of the power unit of the pulsewidth modu-
lator; relays NL and MS operate on the Torquer field to direct the current
flow in either drive or regeneration.
Flyback diodes are used in both configurations to provide a path for
the inductive energy of the Torquer. These flyback diodes are presently
two separate devices, but they also could be combined into one semicon-
ductor switched by a double-pole double-throw relay. However, their low
cost relative to high current relays and their impedance from commutator
requirements favors the two device approach. The main power switching
element SCR, and its associated circuitry are of such a cost that it is
desirable to use a relay to connect it to the PCU in order that it can
be used both in drive and regeneration.
NL must be capable of conducting peak currents up to 400A. To reduce
the size and rating of this relay, a sequence of events during switching
was established so that current never flows through its contacts while
they are open. If this condition can be guaranteed, then a low standoff
voltage contactor (one which can only switch low voltage, but still has
high isolation capability during the non-ionized periods across the con-
tacts) can be used. The difference in volume and weight between a con-
tactor which can interrupt 24 VDC inductive load and a contactor for 200 V
inductive load is approximately 5 to 1.
190
-------
REGENERATION
FLY-BACK DIODE
mm*
^M
mmm
I o , T , J
,. .__-<• — °^ HP n ° r
M3 _...-. , M''
,. roMMUTATIOM
CIRCUIT
=-. v
=- B
[
j>
^
j
1 f !DRIVE
FIELD
J> WINDING i
M4 ^ O— \JUU^T-~ °^^>> M5
CURRbNI SkNbb CURRfcNT
(M4) SENSE COIL
COIL FOR (M5)
RFGENERATION —f- —,— FOR DRIVF
:DRIVE
FLY-BACK
DIODE
ARMATURE
(M5 ENERGIZED)
FIGURE 8.1
POWER ELEMENTS OF THREE RELAY SYSTEM IN DRIVE
-------
REGENERATION
FLY-BACK DIODE
^>
M3
SCR
" in
COMMUTATION
CIRCUIT
1
M4
CURRENT SENSE
(M4)
COIL FOR
REGENERATOR
FIELD
WINDING
FLY-BACK
DIODE
-<
(M3, M4 ENERGIZED)
FIGURE 8.2
POWER ELEMENTS OF THREE RELAY SYSTEM IN
REGENERATION
CURRENT
SENSE COIL
(M5)
FOR DRIVE
ARMATURE
-------
The present EMT system never operates in reverse on the dynamometer.
In an actual vehicle application, reverse driving can be done by gear
shifting or by reversal of the motor terminal polarity. Assuming as a
first case, the motor will always rotate in the same direction, the
armature voltage must retain the same polarity in drive and regeneration,
and therefore the polarity of the field with respect to the armature must
change. In the three relay system, the armature remains permanently wired
to the ground side of the battery, whereas the field is switched. Two
single pole double-throw relays, M. and M,-, are used to invert the field
of the motor, so that it can act as a generator. M. and NL have 24 volt
activating coils used to pull the relays open and closed. In addition,
both coils have a current coil of five turns which carries the load
current of each relay when it is energized. Even if the voltage coil is
deenergized, the relay will be held in the energized position until the
load current drops to a low level. This interlock prevents the relay
contacts from opening while a certain minimum allowable current is still
flowing through the contacts.
The sequence of operation of the three relay system upon the operator
commanding a transfer from drive to regeneration or vice versa is as
follows:
• The duty cycle of the chopper circuitry is reduced to a minimum.
• The power to the logic system of the chopper is turned off,
insuring that commutation ceases and power cannot flow through
the commutation circuit.
0
The power to the 24 V coil on M4 and MS is removed.
• As the current through the five turn coil of M£ or M,. (depending
on direction of the transfer) decays below a minimum level, that
relay drops out.
• When the relay drops out, it disconnects or connects, the 24 volt
coil on M3, establishing the configuration of the power module.
t Contacts on Mj reestablish the 24 volts to M4 or M$ which operate
and transfer the configuration of the series field.
• The power to the logic is reconnected and commutation oscillation
starts.
• Finally the command signal to the chopper is allowed to build-up
to the desired value.
193
-------
8.1.2 Two Relay System
The number of relays in the switching operation can be reduced if
instead of switching the position of the field and SCR,, they are left
in a fixed position and the armature switched. Figure 8.3 shows such a
scheme. SCR, must be located at all times with its cathode at a potential
closer to ground than the anode. In drive it conducts from the high
potential of the battery to ground, and in regeneration from the armature
to ground. Instead of using a relay to switch it from battery potential
to ground level, it can be permanently connected to ground. If in addi-
tion the field is connected in series with SCR,, then its magnetic flux
is always in the same direction in both configurations. The remaining
requirement is to locate the armature in drive so that it can accept bat-
tery power by way of SCR, yet be capable of relocation in regeneration.
A single double-pole, double throw could be used in the application;
however, to insure that there cannot be an opening of the contacts under
load it is safer to retain the two separate relay configurations. This
way transfer in either direction cannot occur until the current has decayed
to low level.
The primary differences between the three and two relay systems are
associated with the PCU's total system interface and the need to sequence
two relays instead of three relays between drive and regeneration. Note
that with the two relay system transition time between drive to regenera-
tion is shorter and vehicle controllability is better.
8.1.3 One Relay System
The diagram of Figure 8.3 shows that the flow of current through the
contacts of M. and M,. are unidirectional. A single SCR or rectifier could
be inserted instead of the four contacts and still retain the circuit
functions if the characteristics of the solid state devices can be mated
to the circuit. The application of semiconductors would decrease trans-
fer switching times, but at some loss in efficiency due to the inherent
conduction and switching losses of the devices.
Figure 8.4 shows a modification of the two relay systems. The M.
relay is replaced by two rectifiers, CRR and CRp. CRR blocks current
flow during drive and conducts during regeneration. CRQ conducts during
194
-------
FIGURE 8.3
TWO RELAY SYSTEM
cn
M5NO
;±FOR REGENERATION
M5NC
FOR REGENERATION
VB
ARMATUREJ J
t^l
DRIVE
ZERO
CURRENT
DETECTOR
COILS
/M
M4
BATTERY
M4NO
FOR DRIVE
FIELD WINDING
M4NC FOR DRIVE
FLY-BACK DIODE FOR
DRIVE AND REGENERATION
t
SCR1
JL f'REGENERATION
Z
o
16
-------
-1- M5NO
FOR REGENERATION
:=:VB
CTi
M5NC 5
FOR REGENERATION
ARMATURE
CRr
SOLID STATE SENSOR
FOR ZERO CURRENT
IN REGENERATION
FIELD WINDING
-l M5 COIL FOR ZERO
CURRENT SENSING.
|N DRIVE
C*n"
CRI:
SCR1
FLYBACK
DIODE
FOR DRIVE
AND
REGENERATION
4 I
I COMMUTATION I
I CIRCUIT |
* ALTERNATE PATH OF COMMUTATION CURRENT
FIGURE 8.4
ONE RELAY SYSTEM
-------
drive and blocks during regeneration. There is no need to use an SCR in
place of CRR since no forward voltage is applied across that path during
regeneration.
The normal operation of this one relay system is as follows. During
drive, SCR, conducts with M5NO closed (M5NC open) and the full battery
voltage is impressed across the motor. Note the blocking voltage of CRD
K
is less than battery voltage. When SCR, is commutated off, current con-
tinues to circulate through the flyback diode and the commutation circuit,
resetting the latter.
Several tests were performed on the one relay system to insure proper
operation before final installation of the all-solid state system. The
following tests were performed:
• Determination of voltages and voltage rise rates. The rate of
change of voltage, -TT-, is an important parameter for sizing the
SCRs, while the diodes are rated in terms of their current
carrying ability and the reverse bias voltages they encounter in
service.
• Determination of the impact of the location of the Torquer
armature inductance on the semiconductor ratings. If the
inductance from the armature is large compared to that of
the series field, then the commutation current through CRR
(or CRD) can be large and the device must be rated at a
higher current level.
• Confirm that the precharge circuit is disconnected during
switching to insure all currents decay to zero during trans-
fer. The replacement of M5 by an SCR hinges on the SCR being
turned off by complete elimination of its current flow. If
this is not done an auxiliary commutation circuit is required.
• Confirmation of switching logic sequences.
8.1.4 Solid State System
SCRR and SCRD were installed in place of the Mr relay and the new
circuit is shown in Figure 8.5. In the drive configuration, SCRD and
SCR, are gated on the current, I,,, flows through the armature, CRD, the
field and SCR,. When SCR, is turned off, current continues to flow through
the free-wheeling diode. During regeneration the current flows through
CRR, the armature, SCRR, the field and SCR, to charge the series field.
197
-------
HUUKt O.b
SOLID STATE SYSTEM
10
00
SCR,
'DRIVE
ARMATURE
— VB
•T«
FLYBACK
DIODE
SCRr
SOLID STATE SENSING
FOR ZERO CURRENT
IN REGENERATION
FIELD
SOLID STATE SENSING
FOR ZERO CURRENT
IN DRIVE
'REGENERATION
--CR,
SCR1
r
COMMUTATION
CIRCUIT
* ALTERNATE PATH OF COMMUTATION CURRENT
-------
When SCR, is turned off, the discharge current continues through CRp into
the battery.
A number of tests were performed to confirm the solid state design
and to gain greater insight into the dynamic requirements of the SCRs ijsed
in this application. The tests included:
• Configuration switching logic characteristics--drive-to-regenera-
tion and regeneration-to-drive
• Blocking voltage stresses on SCRD and SCRR as well as the diodes
of the one relay system
• Measurement of the transfer time for mode switching.
The conversion from the three relay system to the solid state system
increases the speed of switching and 1t becomes extremely important that
the logic characteristics are rapid enough to keep pace with the quicker
system. All the tests conducted indicated that the logic transfer speed
is satisfactory, about 100 milliseconds. This speed is sufficiently fast
so that the operator, when starting to transfer from say drive to regenera-
tion can be insured that logic switching is completed before he is able to
change his mind and command the previous configuration. If transfer was
not completed, a large number of improperly timed sequences could occur.
Figure 8.6 shows an oscillogram of the sequencing from drive to regenera-
tion at a power train speed of 10 mph (1200 rpm for the Torquer). The
upper trace shows the command duty cycle which reduces the duty to cycle
minimum (10 volts) at the start of the sequence. At the same time, the
24 volt SCR gate drive power is reduced to zero. Note that the Torquer
current starts falling immediately. Current continues until the 24 volts
for the commutating SCR gate goes to zero, after which it decays to zero
by virtue of the 20 millisecond motor time constant. After transfer, the
reverse process takes place and drive operation is attained. The
total transfer time is on the order of 100 milliseconds. Previous measure-
ments using the three relay system showed at transfer time of approximately
400 milliseconds, a period longer than operator response time.
199
-------
REGENERATE
120MIL.ISECONDS
DUTY CYCLE COMMAND
SCR GATE DRIVE POWER
VOLTAGE AC ROSS SCRR
TORQUER CURRENT
TIME-
FIGURE 8.6
OPERATION OF SOLID STATE SWITCHING SYSTEM
FROM REGENERATION TO DRIVE
200
-------
8.1.5 Summary of Solid State Switching Activities
An all solid state switching system was developed in a series of steps
starting with a three relay system through two and one relay stages. While
the digital logic is still implemented with low power level pilot relays,
these have no effect on the behavior of the circuits and could be replaced
in the future by low cost integrated circuits.
The solid state system possesses the following advantages:
• Speed of transfer between the configurations is considerably
faster than with relays, with total transfer time less than
operator response time.
t The weight and volume of semiconductors for equal rating is
less than relays.
• The price of SCRs can be reduced by mass production, whereas
the reduction of price for relays is limited by their use of
finite amounts of materials.
t SCR service life is many times longer than that of an equivalent
relay.
On the other hand, the solid state system suffers the following
disadvantages:
• The power dissipation in semiconductors is considerably higher
than relays. Typical diodes and SCRs have 1.5 V forward drop,
whereas relays in the range of current used have approximately
.2 V drop across the contacts.
• The packaging of semiconductors is more critical due to the
need for greater effective cooling. The peak operating tempera-
ture of SCRs is much lower than relays compounding the cooling
problem.
• Semiconductors have considerably lower overload capabilities;
misuse will generally destroy them in circumstances where a
relay would not be damaged.
t Relays can be repaired; semiconductors cannot be salvaged after
damage.
These disadvantages reveal that the use of the all solid state system may
not be the optimum for vehicular use. Instead, further attention should
be paid to the one relay approach which seems to combine the good points
of both switching methods while diminishing the undesirable properties of
each. Some of the advantages in compromising on the one relay approach are:
201
-------
6 The high dissipation semiconductors, SCRn and SCRD, are replaced
by a relay Mg. u K
• Diodes which require no logic to control them replace M^, a device
which must be sequenced in the two and three relay systems.
t MS can be called upon to clear low current level faults, a property
which SCRs do not have (with Me the PCU can be made to malfunction
in regeneration at less than 30 mph and the unit will recover by
itself. With SCRs the PCU will blow a fuse requiring the operator
to stop the vehicle.)
• The Me relay replaces only the higher priced semiconductors, SCRR
and SCRD. K
• Diodes can operate up to 200°C junction temperature, whereas SCRs
are limited to 100°C junction temperature. Diodes are thermally
closer to relays than SCRs.
8.2 EMT Power Train Start-Up Concepts
The breadboard EMT system was started by unclutching the VW engine
from the planetary gear train and turning the engine over by a conventional
VW starter. Once the engine was started, the throttle was advanced and the
clutch engaged. The engine was then brought up to the control speed of the
Speeder.
This approach, while quite satisfactory for the demonstration equip-
ment, is not suited for vehicle use for the following reasons:
• Requires an additional starter motor, starter solenoids, battery
or battery tap.
• Engagement of the engine-gear box is not smooth and engine is apt
to stall. This is particularly troublesome for a lean, quick
relief choked engine needed for reduced starting emissions.
Two approaches for system start-up are suggested; use of the Torquer
as a starting motor and use of the Speeder as a starting motor. Each con-
cept will be described below.
8.2.1 Torquer Starter Motor
In this concept the Torquer-drive wheel connection is broken so the
Torquer can rotate the propeller shaft without moving the car. There is
no clutch between engine and gearing. Using the existing Torquer circuitry
and current transducer, a drive command is given, establishing a torque
202
-------
output of the Torquer. The motor rotates the engine and Speeder. At an
appropriate engine speed, the Speeder tachometer signals the Torquer to go
to normal operation; i.e., full regeneration. The propeller shaft is held
at a low speed and the engine is automatically controlled by the Speeder.
During this period, the Speeder is not operated (the boost PCU is not
turned on), reducing the cranking load of the Torquer. When the operator
wants to move, he engages clutch, making the propeller shaft-rear wheel
connection.
Advantages of the Concept
• Torquer is used as a low speed traction motor
• All cranking torque control is already built into Torquer PCU
• The clutch function can be combined with reversing gears
• The Speeder retains its present configuration and function.
Disadvantages of the Concept
t The engine is always connected to the gearing and Speeder;
design flexibility for gear shifting (see Section 8.3) may
be reduced.
• If the Speeder system fails, the vehicle cannot run on battery
power alone without rotating the Speeder.
8.2.2 Speeder Starter Motor
With the propeller shaft braked, the Speeder is used as a starter
motor. It cranks the engine over to a given speed; once there it reverts
back to its normal generator function. To go forward, the operator re-
leases the propeller shaft and operates the vehicle normally.
Advantages of the Concept
• The need for a clutch between the Torquer and rear wheels is
eliminated.
Disadvantages of the Concept
• The Speeder must be capable of motor operation. Several ways of
doing this are possible:
- use a DC motor/generator as the Speeder with the Torquer PCU
controlling it as a motor, or with a different Speeder PCU
similar to that of the Torquer. The cost of a DC Speeder
would be much greater than an alternator.
203
-------
use the Speeder as an AC motor.
The control of an alternator as a motor from a battery requires an
expensive polyphase inverter and some means of commutating the position
of the rotor.
The most attractive method of combining the EMT with engine starting
is to use the DC Torquer with a slightly modified form (engine speed sens-
ing and mode transfer control) of the Torquer PCD.
8.3 Override Options to the EMT System
A drawback of the EMT system in its breadboard form is its inability
to provide sustained low speed (lower than mode transition speed), high
power (greater than the Mode One engine power). This section suggests
several options which could be included in an advanced EMT system. They
are restricted to overrides during Mode One and would be either manually
or automatically controlled.
8.3.1 Option A
In this option, the engine speed increases at relatively constant
engine torque. This is done by increasing the engine throttle while simul-
taneously increasing the engine control speed level of the Speeder. As
the Speeder and engine increase their power output, the incremental power
is delivered to the Torquer for traction. The limiting factor is the
design speed of the Speeder. In the present system, there is very little
reserve Speeder speed at very low road speeds; near mode transition, there
is almost the full Speeder design speed available and the engine output
can be approximately doubled.
8.3.2 Option B
This option lets the engine torque increase at constant engine speed.
Increasing the throttle, increases the torque which the Speeder must
accept; however, it also increases the torque delivered directly to the
propeller shaft. The former effect means more power can be converted by
the Speeder for use in the Torquer. The limit of usable engine power is
set by the sustained torque capability of the Speeder PCU. At speeds near
mode transition where the Speeder voltage is low, the increased current
output can be limited by the lack of storage capacity in the boost inductor.
204
-------
8.3.3 Option C
In this option the combined techniques of options A and B arc used.
It is estimated that the combined technique could increase the factor of
engine output from the normal Mode One level by about 2.5 to 3.
8.3.4 Option D
This option relies on a gear ratio changing between the engine and
the rear wheels; in effect, down-shifting the transmission with an asso-
ciated engine throttle increase. One method of doing this is to manually
instruct the transmission to go into Mode Two (shut the Speeder off-line)
while simultaneously changing the gear ratio between the engine shaft and
the planetary input or the planetary output and propeller shaft. For the
present system this would be equivalent to shifting from 4th gear (the
approximate overall gear ratio of the breadboard equipment) to something
between 1st and 2nd gear. Near mode transition, the engine output could
approach its maximum level without overloading or overspeeding the Speeder.
8.3.5 Conclusions
Of the options offered, A, B and C will.not require substantial re-
design of the EMT system but rather modification of existing Speeder and
engine throttle subsystems. On the other hand, these options will not
allow the full power output of the engine to be utilized without excessive
weight and cost penalties to the Speeder system. Option D, while requiring
a secondary gear ratio allows full engine output to be developed near mode
transition. Further tradeoffs must be made to establish on one hand the
degree and frequency of sustained power demand in Mode One and on the other
hand the detailed design implications of each option.
205
-------
9.0 EFFECT OF SELECTED PARAMETERS ON EXHAUST EMISSIONS FROM AN INTERNAL
COMBUSTION ENGINE
9.1 Introduction
This study was undertaken to determine the effects of various engine
parameters during the start-up period of engine operation. In particular,
emphasis was placed on defining the boundaries of choke scheduling, timing
and power so that the engine would "deliver stable power with minimum
emissions. The EMT system is a particularly attractive system to conduct
the experiments as the Speeder load control completely frees the engine
from the driveability requirements of"normal automotive operation.
The work performed falls into three categories. The first deals
with a series of experiments in which the cold start emissions of the
engine were measured under a variety of controlled start conditions.
Measurements were made using both the proportional sampling technique
and continuous on-line concentrations.
The second data set defines engine operation beyond the cold period.
Particularly lean engine operation and the effect of timing and power
level were explored. Data was again gathered by both "bag" and concentra-
tion measurements.
Finally, some data was experimentally obtained on the performance of
a three-component catalyst operating under steady hot emission load.
9.2 Determination of Engine Emissions During Cold and Hot Operation
9.2.1 Experimental Equipment and Procedures
The test engine used was a 1968 VW 1600 cc engine equipped with in-
take port fuel injection. It is the same engine used in the dynamometer
work of Section 5 and was operated into the power train described in the
earlier sections. The fuel injection control system was modified in two
ways. First, the injection pulse duration circuitry was adapted so that
the mixture ratio of the engine could be varied over a wide range from
flood to lean misfire. These modifications affected a continuously vari-
able metering jet capability to the engine carburetion. Second, the
thermal transducer controlling the choke operation was removed and a
206
-------
manually controlled signal was added in its place so that the "choke"
could be varied over a wide range of settings and time schedules.
The engine was instrumented as follows:
Engine air flow - a Meriam Laminar flowmeter was used to monitor the
engine intake air flow. The pressure drop across the flowmeter was
monitored by an inclined manometer and recorded using a pressure trans-
ducer.
Fuel flow - a turbine meter calibrated in-situ monitored the fuel
flow.
Manifold pressure - the manifold pressure was indicated by a mercury
manometer and pressure transducer.
Engine temperatures - two thermocouples were placed in the exhaust
manifold about one inch from the manifold flange on the number 3 and 4
cylinders. Additional thermocouples sensed the temperature of; the outer
portion of one cylinder lead, the crankcase oil and the mid-point of the
muffler.
Engine speed - engine speed was monitored by a tachometer-generator
driven off the engine cooling fan. This tachometer is currently used as
a primary sensor for the Speeder control system.
Engine timing and control - the engine timing was adjusted by two
means. First, the distributor was manually rotated from the manufacturer's
specified position (0° TDC) with engine off. Additional timing control
was affected by using the vacuum advance diaphram on the distributor in
conjunction with an external vacuum pump. All timing was checked with
an electronic stroboscopic tachometer.
Emission measurements were made using two sets of equipment:
• Continuous concentration monitoring of the exhaust was made by
drawing a sample from the mid-point of the muffler. The sample
line was fed into the following instruments:
Beckman Model 315A NDIR CO Analyzer
Beckman Model 315A NDIR NO Analyzer
Beckman Model 400 FID Hydrocarbon Analyzer
Beckman Process Oxygen Analyzer
207
-------
• Proportional sampling (1972 HEW test equipment) was performed by
Scott Research Laboratories using their equipment. The technique
samples and measures the CO, NO, N02 and FID hydrocarbons in a bag
diluted with ambient air. Additional on-stream records of the NDIR
concentrations of CO, C02, NO and hydrocarbons were taken. Bag
measurements were performed on approximately 50% of the cold start
tests and 30% of the hot tests.
Power delivered by the engine was determined by recording the voltage
and current of the Speeder rectifier output. The values of the inefficien-
cies of the Speeder and the gear box were derived from the previous experi-
mental work (Table 5.2 and Figure 7.2) and were factored into the computa-
tions to reflect the engine shaft output power.
9.2.2 Cold Start Test Results
The data gathered during this portion of the study is summarized in
Tables 9.1 and 9.2. A total of 14 cold starts were performed.
The mass rates in Tables 9.1, 9.2 and 9.3 of the next section were
computed from measured steady state or averaged pollutant concentrations and
engine air intake flows. Intake air mass flows were corrected for the addi-
tion of fuel to compute the exhaust mass flow rates. The test numbers of Table
9.1 refer to consecutive 1^ min. intervals during which samples were taken.
Table 9.la compares the bag measurements to the computed ("integrated") pol-
lutant mass rates for the samples taken simultaneously. Good agreement is
observed in hydrocarbon mass rates, but the integrated CO and NO rates are
consistently lower than those obtained by the constant volume bag technique.
The discrepancy in CO mass rates is especially high. An investigation of
the observed differences revealed that most of the hot test bag measurements
of CO and NO were recorded at the least sensitive portion of the NDIR instru-
ments (lower 5% of scale); this can easily be the cause of the observed error
during hot work. However, the CO error should not occur to the much higher
levels in CO.
For the purpose of this study where trends are more important than
absolute values, the integrated mass rate results presented in Tables 9.1
and 9,3 were used to establish relationships between them and engine opera-
tion parameters. The choice was dictated by the fact that the integrated
mass rates represent a more complete set of data.
208
-------
TABLE 9.1 COLD
Test Nos.
C-l.l
1.2
1.3
C-2.1
2.2
2.3
C-3.1
3.2
3.3
C-4.1
4.2
4.3
C-5.1
5.2
5.3
C-6.1
6.2
6.3
C-7.1
7.2
C-8.1
8.2
8.3
C-9.1
9.2
9.3
C-10.1
10.2
10.3
C-ll.l
11.2
11.3
C-12.1
12.2
12.3
C-13.1
13.2
13.3
C-14.1
14.2
14.3
Initial Conditions
Cold
Soak
Period
(Hours)
2
2
Over-
night
3
3
Over-
night
7
Over-
night
3
2
2
Week-
end
4-1/2
3
i Crank-' ;
1 case |
j Temp. ; Timing :
: 70 ; 10°retard/ :
Full vac.
adv.
: 70 Normal VW
; 70 10°retard,
' VW Vac. adv.
1 80 ! 10°retard
: Full vac.
; adv.
• \ .. . ,
80 10°retard
i Full vac.
I i adv. j
; 60 ; 10°retard :
; : Full vac.
| ; adv.
' 75 i 0°retard \
> Full vac. •
• adv.
j 70 I Normal VW
t •
: 75 Normal VW
• • i
i :
j 80 i 0°retard ;
i : Full vac.
- j : j
i retard \
| 75 j 0°retard |
i • Full vac. j
i ; adv. j
i • j t •
! 70 ' Normal VW i
i i i
i ' '
i 75 " i 10° retard |
i • Full vac. i
i ' adv. ;
i 70 ; o°retard j
j Full vac. j
i j adv. j
Throttle
Setting
1.69
1.55
1.55
1.35
1.11
1.11
1.11
1.11
1.55
1.35
1.35 \
1.35 •
1.35 i
1.35 i
Choke
Manual
Control
VW
VW
Manual
Control
Manual
Control
Manual
Control
Manual
Control
VW
VW
Manual
Control
Manual
Control
VW
Manual
Control
Manual
Control
Exhaust
(Avg.@l.
A . S
HC
Gms/Min.
4.18
1.65
1.03
5.08
1.77
1.18
6.2
1.42
0.66
2.65
0.49
0.38
2.95
0.83
0.60
2.10
1.03
0.58 ;
3.04
1 . 30
4.00
3.06 :
1.52
2.06
1.38
2.12
0.59
0.52 ;
2 . 25
1.34 i
0.91
5.78 ;
1 . 62 ;
1 . 10 :
3.87
1.09 i
0.49
2.48 i
0.81 i
0.72 :
Composition
5Min. intervals for
M-irmt-psI
CO
Gms/Min.
6.45 ;
1 . 65 :
1.53 ;
38.1
2.07 '
1.30
37.6 :
1.85 i
1.40 \
9.1 '•
1.15 i
1.11
8 . 30 i
1.48 \
1.32 i
16.0 :
1.18 ;
0.95 I
32.8 :
1.02
53.4
23.1
5.1
6.54
1.71
23.80 :
4.84
1.41
9.8 :
1.4 !
1 . 25
57.0 •
4.05
1.10 ;
9.45 '
1.89
1.35 ;
5.6 •
0.9 i
1.03 i
NOx
Gms/Min.;
1.14
0.47
0.49
2.14 I
5.38 i
5.38 1-
1.02 j
2.62 I
2.54 I
0.48 !
0.41 1
0.30 !
0.48 '
0.73 ;
0.73 :
0.22 \
0.30 1
0.49 !
0.45 ;
1.03 \
0.12 i
1.70 .
2.36
4.58
6.42
0.38
0.90
1.28
1.24
4.46 .
4.42
0.33
3.20 '
4.32
1.11 '
1.22
1.84 :
1.12
1 . 06 :
1.41 i
°2
5
5
5
0.6
1.8
2.4
1.2
2.3
2.5
4.0
4.5
5.0
4.5
4.5
4.5
2
5
5
5
4.5
1.3
1.3
2.0
1.2
1.6
2.2
1.2
3.3
3.5
4.5
3.0
3.2
1.2
1.9
2.4
4
4
3.5
4.0
4.6
4.6
-------
START TESTS OF 1600 cc.VW
Average
Intake
Air
Flow
(cfm)
35.6
35.8
35.8
31.4
31.4
JO. 8
30.8
32.2
31.2
24.0
25.0
25.0
24.0
32.0
32.0
16.5
19.2
23.8
24.5
30.8
17.0
24.0
23.0
32.4
32.4
32.4
24.0
24.0
26.7
33.0
3£.5
35.0
ni f
24.5
24.4
39.4
41.8
37.2
23.0
24.5
26.0
•••••M
tngine
; Exhaust
..; Gas Temp .
U310 Final
[1180 @ 20sec.
:1260 Final
;1130 @ 45 sec
1 1320 Final
;1190 @ 45sec.
;1300 Final
il!70 @ 30sec.
•1300 Final
;i!70 @ 55sec.
1275 Final
;1150 @ 70sec.
1250 in 4min.
i;il20 in 50 ;
1 sec. ;
^1220 Final \
1100 @ 80sec.
•1250 Final ]
1120 @55sec i
i!270 Final }
:1140 @48sec. \
'1260 Final ;
11135 @45sec.
I
• i ? "}("> vi a i
;1100 @90sec.
i!380 Final
1240 @55sec.
:i200 Final i
:1080 @35sec. i
! i
Operating
i
RPM
1800
1
1800
1
1800
1800
1
Oscil-
lation
1800
1400
1400
1800
j .
1400 ;
1800 ! ^
1800 :
1800 :
1800 i
1
1800
1 1
1 ftOO
1 i
1400
1800
1800
Conditions
Alternator
Output
(kw)
2.06
3.7
5.5
6.7
5
5
5.5
2.55
3.22
5.5
<2
<2
^4.0
8.15
7.40
7.05
2.66
7 . 4 max .
(4)
L 0
4.8
4.6 ;
6 . 2 (max)
5.7
2.8
3.2 j
! Manif.
• Press
i
6.3
T
8.3
12.3
12.3
12.5
5.8
5.8
<12.5
^12.5
9.5
@60|sec
13.3@2min.
9.0
8.86
8.6
13.1
10.2
3 . 9max
6.5
i ? n
11.85
11.85
2.0 max
5.8final
12.2
12.2
12.2
Remarks
Choke completely off in 60 seconds.
Normal VW start with fixed throttle
Normal VW with fixed throttle,
10° retard, 6 seconds for engine
to start.
Engine stumble at 25 & 60 seconds
due to under choking leaned
during, cold start).
Increased throttle @ 28sec., engine
too lean, did not re-establish control
until 45 sec., speed & emission excur.
Opened throttle in stages @ 1:40,3:00,
3:30. Excessive enrichment during
start (3sec.)No regulation till 3:40.
Increased throttle at 30sec. through
60sec. ;dropped choke @ 45sec.to fully
off; engine stable in approx.60sec.
Normal VW start, increase throttle
after 30 sec., 1400 rpm till 35sec.
(1 minute intervals).
1 minute intervals
VW choke schedule for 35sec., then
rapid decrease to full off in 1:00 mil
1 minute intervals.
After 4-5sec. cranking, leaned down for
lOsec. .retarded spark at 30sec., inc.
throttle @50sec. .reduced thr.HOsec.
leaned during tull test.
Normal VW start.
Engine cranked 10 seconds.
Lean down for 30sec. then retard
spark 10° & adv. thr . (Eng.out-of-cont .
for 40sec.) Cont .decreasing thr. &
leaning.
Rapid leaning. Most choke relief
after 20 seconds.
209
-------
Test No.
C-8.1
8.2
8.3
C-9.1
9.2
9.3
C-10.1
10.2
10.3
C-ll
11
11
C-12.1
12.2
12.3
C-13.1
13.2
13.3
C-14.1
14.2
14.3
TABLE 9. la
COMPARISON OF CVS AND INTEGRATION PROCEDURE COLD START EMISSION VALUES
grams/min
CVS Method
HC
2.57
1.93
1.36
6.24
1.85
2.19
2.34
0.50
0.50
1.76
0.72
0.94
5.80
1.72
1.42
3.14
1.20
0.58
48.37
24.19
16.55
86.56
10.18
5.09
54.45
2.42
2.42
8.47
2.42
2.42
125.21
2.46
1.23
8.59
3.68
3.68
2.70
1.13
0.88
11.05
2.46
2.46
N0x_
0.31
2.01
5.00
2.41
7.57
9.03
0.84
1.21
1.69
1.75
4.08
4.27
0.75
6.01
5.06
1.77
1.92
2.82
1.25
1.11
Integration Procedure
HC
4.00
3.06
1.52
2.06
1.38
2.12
0.59
0.52
2.25
1.34
0.91
5.78
1.62
1.10
3.87
1.09
0.49
2.48
0.81
0.72
CO
53.4
23.1
5.1
6.54
1.71
23.80
4.84
1.41
9.80
1.40
1.25
57.00
4.05
1.10
9.45
1.89
1.35
5.60
0.90
1.03
NOX
0.12
1.70
2.36
4.58
6.42
0.38
0.90
1.28
1.24
4.46
4.42
0.33
3.20
4.32
1.11
1.22
1.84
1.12
1.06
1.41
-------
TABLE 9.2
TEST
NUMBER*
C-2 & C-9
C-3
C-12
C-8
ENGINE PARAMETERS
Normal VW choke and timing
High Power Settings
Normal VW choke
Mech an ica ^Retarded 10°
Normal VW choke and timing
Medium Power Setting
Normal VW choke and timing
Low Power initially, scheduled
throttle** after 35 sec.
POWER
END OF
3 MIN
(HP)
15-16
13
12
11
C-10
C-l
C-4
C-5
C-6
C-7
C-ll
C-13
Normal VW choke for 35 sec only 9
Full Vacuum Retard
Rapid de-choking after 35 sec
off choke in 1 min.
Controlled VW (EMT) 10
High Throttle.10° Mech. Ret.
Lean start***, 60 sec. de-
choking schedule
EMT - Low Throttle.10° Mech. 9
Retard, Lean start, 60 sec.
de-choking
EMT - Same as C-4 but 13
scheduled throttle
EMT - Low throttle, 10° Mech. Ret. 8
Rich start
Rapid choke relief, stepwise throttle
increase
EMT - Low throttle, 0° Retard
scheduled choke and throttle
through first 60 sec.
EMT - High throttle
0° Retard initially
10° Retard after 30 sec
(see Table 1 for choke schedule)
EMT - High throttle, 10° Ret.
initially - 10° additional retard
after 30 sec (see Table 1 for
schedules)
C-14 EMT - Low throttle, 0° retard
Rapid de-choking (20 sec)
11
16
15
TOTAL POLLUTANTS
HC CO NO
REMARKS
10.27 60.12 11.28 Good start-reproduci-
ble
11.43 59.17 5.46 Rough start
11.10 91.58 5.30 Rough start
8.58 89.11 7.32 Maximum RPM first
35 sec. 1400
3.23 30.25 2.56 Good start
8.74 12.15 2.41 Good start
4.71 15.37 1.34 Rough engine opera-
tion first 60 sec.
5.67 14.57 1.81 Rough up to 45 sec.
No speed
4.70 25.77 0.78 . control during
entire test (RPM
1400 rpm) ,
6.51 50.73 2.22 Good start
5.39 16.8 8.55 Good start
No speed
7.44 17.01 3.50 control for 40 sec
(1400 RPM)
4.94 9.75 3.27 Good start
**
***
Corresponds to Table 9.1
Implied scheduled increase in m^ifold pressure
"Lean" implies leaner than normal VW
211
-------
Typical pollutant concentration signatures during the cold period
are shown in Figures 9.1 and 9.2. In Figure 9.1, the engine was started
with the normal VW choke and timing schedule. The throttle was set to
give approximately 9 hp after three minutes of operation. Figure 9.2
shows a cold start test at the same three-minute power level but with a
much more rapid choke relief. Figures 9.3 and 9.4 show temperature and
air-fuel ratio variations for the respective tests.
Several observations can be made. A low air-fuel ratio (< 10) is
necessary for engine start. This very rich mixture results in a high
hydrocarbon and CO output during the initial period of operation (^ one
minute). Second, the engine reaches a quasi-steady state within three
minutes after ignition. Third, rapid choke relief dramatically reduces
both the concentration magnitude and duration of hydrocarbon and CO
emissions. NOX concentrations, however, start to buildup earlier in the
operating cycle.
Figure 9.5 demonstrates the effect of rapid choke relief on total
emissions during the first three minutes after engine start. All major
pollutants are decreased by rapid choke relief when contrasted to the
normally choked engine. The degree of reduction is a function of the
particular pollutant. CO and hydrocarbon show the highest and inter-
mediate percent reductions; NO reduction is considerably less. CO and
HC reductions are somewhat independent of engine power level, suggesting
they are associated with the earliest period of engine operation where
the engine is stabilizing its internal temperature structure. NO levels
on the other hand are more related to the three-minute power level.
With increasing manifold pressure (increased power) the NOX concentration
and volume flow rate increase once the engine has passed through its
initial starting period.
Figure 9.6 indicates the effect of engine timing on total pollutant
levels. Better than 50% reduction in NOX is observed with 10° retard at
both power levels. However, spark retard appears to have an adverse
effect on hydrocarbon emissions; the effect increasing with power level.
The engine is harder to start with a retarded spark and the incremental
HC levels may be due to additional fuel injected into the engine during
cranking.
212
-------
ro
CO
LLJ
U
O
U
NORMAL VW 1600 COLD START
EMT-11 HP (TEST C-8)
30 60
90 120 150
TIME - SECONDS
18,000
15,000 -i4000
12,000 — -
9000
6000
3000
Q-
Q_
Z
O
CO
CXL
<
u
O
Qi
Q
3200
2400
1600
800
Q_
Q_
X
O
z
180 210 240
FIGURE 9.1
-------
CONTROLLED VW - 1600
COLD START
EMT - 9 HP (TEST C-14)
10
9
K- 8
Z
ULJ
U 7
UJ
a.
' 6
O
">
4
3
2
1
C
i
A
1
i
!
'P
/
/
/
\ /
u
J
hi
*
i
>
\
i p
A
\
\
\
i
\
\
\
\
\
\
\
\
\
•
\co ^~
\y
*
/
^^•^^•••i^B • *•
X
s
• —
A
^'
•*- '
NOX
HC
30 60 90 120 150 180
12,000
11,000
10,000
9,000 ^
i^
8, 000 5
a.
a.
7,000 '
o->
z
6,000 0 _
Of
5,000 5 -
Q£
4, 000 §
3, 000
2,000
1,000
1500
1400
1300
1200
1100
1000 |
a.
900 'x
O
800 Z
700
600
500
400
TIME - SECONDS
FIGURE 9.2
214
-------
1300
NORMAL VW 1600 COLD START
EMT - 11 HP (TEST C-8)
No. 3 CYLINDER TEMPERATURE
No. 4 CYLINDER TEMPERATURE
AIR- FUEL RATIO
MUFFLER TEMPERATURE
120 150
TIME - SECONDS
180
210
240
FIGURE 9.3
215
-------
CONTROLLED VW-1600 COLD START
EMT-9 HP (TEST C-14)
1200
No. 3 CYLINDER TEMPERATURE
No. 4 CYLINDER TEMPERATURE
AIR-FUEL RATIO
MUFFLER TEMPERATURE
60 90 120
TIME - SECONDS
g
i
14 3
I
ae
12
10
FIGURE 9.4
216
-------
EFFECT OF CHOKE SCHEDULE ON
3 MINUTE POLLUTANT TOTALS
I**
12
1
10
oo
z
o
to
5
"x 8
§
Q
Z
^
I «
CO
-
2
0
16 HP AT 3MIN
10 HP AT 3MIN
• •• ^m
-«* — -
EMT VW EMT VW
HC NOX
"™" ** * *
EMT VW
(•4W
120
100
^
^
2
80 J>
Z
O
LO
I
tu
60 0
_J
t—
O
t—
40
20
0
CO
FIGURE 9.5
217
-------
CO
no
00
c:
73
vo
10
o
z
CO
z
o
CO
X
o
Q
z
<
CO
z
o
CO
u
O
Q
£
—i
<
O
EFFECT OF ENGINE TIMING ON
3 MINUTE POLLUTANT TOTALS
SOLID L
1
~9
HP
1
NE 10
-15.5
HP
0 RETARD WIT
t
~9
HP
H RESP
-15.5
HP
ECT TO DASHED LINE
**• 15 5
~ 0 \~J--J
LJD
HP . _ .
. .
100
80
60
CO
1
O
I
CO
z
o
CO
CO
40 O
20
LOW HIGH
HC
LOW HIGH
LOW HIGH
CO
POWER
-------
Several conclusions can be drawn from the cold start test work:
t The internal combustion engine tested exhibits a period of
operation after ignition where its temperature, mixture ratio
and emission signature vary very rapidly.
• This period is of approximately three minutes duration.
• The cold period can be broadly broken into two portions; the CO
and HC portion and the NOx portion. The former occurs immediately
after ignition and lasts for 30 to 60 seconds. During this period
the major contribution to the cold start HC and CO emissions
results. The NOx portion builds as the HC and CO portion falls
off. While the HC and CO mass totals are relatively unaffected
by power level, the NOx contribution is clearly related to power
level.
t Emissions during the cold start period can be greatly reduced by
controlled engine operation. Simultaneous rapid dechoking and
throttle increase may be the best starting procedure.
t Spark retard reduces the mass emissions during the cold periov
Reduction of NOx levels is particularly sensitive to timing; the
effect increases with increasing power level.
t Minimum emission^ are obtained from procedures such as followed
in tests C-10, C-ll and C-14 of Tables 9.1 and 9.2.
9.2.3 Hot Tests Results
A complete summary of hot engine test results is given in Table 9.3.
Thirteen tests were performed varying in duration from ten to thrity
minutes. During each test a number of engine parameters were varied and
their effect on pollutant mass rates was determined. Tables 9.4, 9.5 and
9.6 group a portion of the data of Table 9.3 in a way such that the effects
of various engine parameters on emissions are isolated.
The data of Table 9.4 indicates the effect of air-fuel ratio on nujor
pollutant mass rates at constant power, speed and engine timing. Two power
settings and two values of ignition timing are examined. The data of Table
9.4 indicates that at both power settings and both spark timings minimum
pollutant mass rates are observed at leaner than stoichiometric air-fuel
ratios. It is also shown that at constant spark timing the higher the
power setting the leaner the fuel mixture should be to keep emissions low.
Increased spark retard requires a decrease of the air-fuel ratio for con-
stant power operation. In general, even a five percent change in the
219
-------
TABLE 9.3 HOT TESTS (1968 VW)
Test
Nos.
1-1
1-2
1-3
1-4
1-5
1-6
1-7
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
2-9
Enqine Parameters
Air/
Fuel Timing
Ratio BTDC
16.1 10°
16.1 20°
17.2 20°
14.5 20°
17.8 20°
19.0 20°
17.7 20°
19.1 20°
18.3 10°
16.9 20°
15.2 20°
16.8 20°
16.6 20°
17.3 20°
17.7 20°
19.8 20°
I
3-1 1 12.1 30°
3-2
3-3
3-4
3-5
3-6
3-7
3-8
3-9
3-10
3-11
4-1
4-2
4-3
4-4
4-5
4-6
4-7
4-8
4-9
4-10
4-11
12.3 30°
13.7 30°
16.2 30°
18.0 30°
19.4 30°
16.8 30°
16.9 20°
16.9 22°
17.0 30°
16.9 30°
16.1 20°
14.9 20°
14.3 20°
13.0 20°
13.3 20°
13.3 10°
13.3 20°
15.1 20°
15.6 20°
15.3 20°
15.2 20°
Manifold
Pressure
"Hg
8.4
8.4
6.5
6.7
6.7
6.7
6.7
6.2
6.0
3.0
3.0
0.9 .
0.9
0.9
0.9
0.9
8.2
8.3
8.4
8.4
8.4
8.4
8.5
8.4
8.4
8.4
0.90
8.3 .
13.4
13.4
13.4
13.4
13.4
13.4
13:4
10.4
9.3
RPM
1800
1800
1800
1800
1800
1800
1800
1800
1780
1800
1800+
1800+
"1800+
1800+
1800+
1800
1800
1800
1800 >
1800
1800
1800 '
1800
1800
1800
1800
1800+
1800
1800
1800
1800
1800
1800
1800
.-1800
1700
1510
1330
Exhaust Composition -
Concentration
HC CO
as C,J'
PPM)J (PPM)
FID
190 950
360 1400
250 1100
370 1750
220 1100
600 1420
200 1110
300 1150
120 f 750
170 1000
330 2550
200 1100 _
230 1130
150 1000
160 1050
330 1150
1400 4150
1200 2500
750 3200
400 1250
350 1200
1200 1700
400 1250
230 1100
250 1200
360 1250
540 1200
350 1350
250 1200
250 1350
710 11000
450 4000
200 3000
450 4400
200 1100
160 1000
150 1000
450 1400
NO
(PPM)
980
2000
550
2300
300
170
900
750
300
1100
2400
2000-
2350
750 -
350
210
1250
2000
3250
2150
600
150
1350
650
800
1300
3700
1600
900
1100
1300
1400
800
1450
900
600
300
150
°2
(«'
3
2.5
3.5
1.8
4.5
6.0
4.0
4.5
5.0
5.0
5.0 .
.5.6
0.5
1.2
2.5
3.5
4.6
6.0
3.70
3.70
3.70
4.00
3.20
2.50
2.0
1.5
0.5
0.8
0.7
0.8
2.2
3.0
3.7 .'
5.0
Emissions -
Mass Rate
HC
gms/
m1n
0.28
0.53
0.42
0 .60
0.35
0.97
0.32
0.53
0.20
0.33
0.60
, 0.43.
0.51
0.33
"0.35
0.73
2.10
1.80
1.12
0.60
0.52
1.80
0.60
0.34
0.37
0.54
1.18
0.54
0.26
0.27
0.76^
0.48
0.21
.0.48
0.21
0.17
0.15
0.43
CO
gms/
min
0.94
1.38
1.23
1.90
1.20
1.55
1.20
1.36
6.7.7
1.21
3.42
.1.61
1.68
1.48
1.56
1.70
4.20
2.52
3.23
1.26
1.21
1.72
1.26
I'.ll
1.21
1.26
1.76
1.39
,0.85
0.98
8.00
2.90
2.20
3.20
0.80
0.73
0.70
0.89
NO
gms/
min
1:60
3.26
1.02
4.20
0.54
0.305
1.61
1.46
0.55
2.43
5.29
4.80-
5.73
1.84
0.85
0.51
2.08
3.32
5.40
3.58
1.00
0.25
2.24
1.08
1.33
2.16
8.94
2.70
1.05
-1.31
1.55
1.67
0.95
1.73
1.07
0.71
0.35
0.16
Engine
Performance
Intake Alter-
Air nator
Flow (Kw)
(CFM)
30 2.52
30 4.0
34
33
33 2.91
33 1.92
33 3.60
35.8 3.70
34.0 1.9
40.4 5.9
40.5 8.45
44 ..3 9.08
45.0
45 6.96
45 5.68
45 4.54
30.6 5.5
30.6 7.5
30.6 7.4
30.6 5.6
30.6 4.5
30.6 2.4
30.6 5.75
30.6 4.55
30.6
30.6
44.5 11.3
31.2 5.5
21.5 2.44
21.9 2.96
21.9 3.80
21.9 3.53
21.9 2.44
21.9 3.70
21.9 .2.61
21.9 2.28
20.9 2.23
20.0 0.7
Exhaust
Gas Temp
TO
1420
1330
1380
1330
1430
1400
o
CM
CVJ
1360
1260
1300
1250
1320
1280
1350
1320
1350
1430
1350
1300
1300
-------
5-1 17.1 20°
5-2 17.3 20°
5-3 16.7 20°
5-4 17.8 20°
5-5 19.3 20°
5-6 20.0 20°
5-7 16.4 20°
5-8 15.1 20°
5-9 12.4 20°
6-1 16.2 20°
6-2 16.1 10°
6-3 15.6 10°
6-4 15.6 20°
6-5 16.4 20°
6-6 15.1 20°
6-7 18.1 20°
6-8 18.4 10°
6-9 13.4 10°
6-10 14.4 10°
6-11 15.7 v 10°
7-1 15.0 30°
8-1 14.8 30°
8-2 15.6 30°
8-3 15.6 20°
9-1 16.6 20°
9-2 17.7 20°
9-3 17.5 20°
9-4 16.8 20°
10-1 14.1 30°
10-2 13.8 30°
10-3 13.8 30°
10-4 14.6 20°
11-1 15.5 10°
12-1 17.1 20°
12-2 18.0 20°
12-3 16.8 20°
13-1
13-2 - 20°
13-3 - 20°
0,9
0.9
0.9
8.7
5.9
2.7
9.9
12.0
13.6
9.8
9.0
10.1
10.1
8.25
9.45
6.20
6.30
8.6
7.8
6.5
13.1
13.2
8.7
3.0
10.2
10.1
7.9
0.9
11.85
14.1
14.1
0.85
6.4
7.5
6.2
0.95
10.2
6.0
8.5
J.800
2000
1800
1800
1800
1800
1800
1800
1800
^1800
1700
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
2100
1500
1800
420
350
290
170
220
550
120
170
880
260
120
140
280
250
380
190
110
550
260
150
700
640
590
520
350
250
250
270
790
740
705
600
140
270
185
230
470
580
450
1250
1300
1350
1100
1250
1750
1150
1300
3500
1150
700
1000
1550
1250
2300
1130
750
17000
6000
1250
1400
1425
1275
1750
1175
1100
1130
1140
1370
1525
1525
2450
1000
1500
1100
1450
1400
2000
1320
2200-
2200
1200
400
250
250
180
660
1200
710
420
950
1850
1750
2250
800
280
1250
1600
1450
2750
2600
3400
3600
1300
475
1130
2350
3250
2340
2375
3810
800
920
395
1970
2000
2100
1850
3.2
3.0
4.0
4.2
5.3
6.0
3.2
2.0
0.4
4.0
3.8
2.0
2.0
2.8
1.3
4.2
4.2
"'0.4
0.6
1.7
2.2
2.1
2.2
1.8
3.5 '
4.0
3.5
3.5
2.4
2.0
1.8
1.5
3.4
4.2
5.1
4.2
1.7
1.4
2.2
0.94
0.86
0.62
0.25 .'
0.39
1.14 .
0.16
0.20
0.95-
0.36
0.17
0.19
0.38
0.37
0.53
0.32
0.19
0.81
0.40
0.25
0.79
0.72
0.86
1.05
0.49
0.34
0.40
0.59
0.95
0.73
0.69
1.29
0.23
0.42
0.31
0.50
0.72
0.74
0.64
1.87
2.14
1.95
1.10
1.47 .
2.43
1.06
1.04
25.2
1.07,!
0.65
0.93
1.44
1.24
2.20
1.29
0.87
16.8
6.23
1.42
1.07
1.08
1.25
2.40
1.10
1.02
1.21
1.69
1.10
1.00
1.00
3.55
1.13
1.58
1.25
2.12
1.44
1.60
1.27
5.30
5.96
2.86
0.66
0.49
0.57
0.27
0.87
1.42
1.10
0.64
1.44
2.80
,2.86
/
3.50
1.50
0.52
2.04
2.70
2 . 70
3.44
3.24
5.50
.8.10
2.00
0.72
1.98
5.72
4.30
2.54
2.56
9.10
1.48 .
1.60
0.74
4.73
3.40
2.96
2.94
'45.4 9.3
50.0 9.9
43.8 8.52
30.4 3.75
35.8 3.74
42.2 3.75
28.0 3.75
24.2 3.75
21.8 3.75
28.2 3.63
28.0 2.28
28.0 3.63
28.0 5.5
30.2 6.23
28.8 6.23
34.6 6.23
34.5 2.96
30.0 6.02
31.5 6.23
34.4 6.23
' 23.0 3.86
23.0 3.86
29.7 6.85
41.4 10.80
28.4 4.10
28.0 3.96
32.4 5.6
44.8 10.6
24.4 4.55 .
20.0 3.45
20.0 3.37
43.9 12.5
34.0 4.86
31.9 5.00?
34.3 4.45?
44.2 9.75
31.2 4.28
26.0 4.93
29.2 4.0
1350
1380
1350
1320
1420
1440
1350
1360
1460
1250
1250
1270
1340
1300
1300
1310
1310
1240
1240
1240
1370
1400
1300
1310
1320
1400
1340
1350
-------
TABLE 9.4
EFFECT OF AIR-FUEL RATIOS ON POLLUTANT MASS RATE
A. 10.5 hp; 1800 rpm; 20° BTDC
Oxygen
Exhaust in Pollutant Mass Rates (grams/min)
Test Nos. Air-Fuel (Percent) HC CO NOX (as NQ2)
5-9 12.4 0.4 0.95 2.52 1.42
5-8 15.1 2.0 0.20 1.04 0.87
*5-7 16.4 3.2 0.16 1.06 0.27
5-4 17.8 4.2 0.25 1.10 0.66
5-5 19.3 5.3 0.39 1.47 0.49
5-6 20.0 6.0 1.14 2.43 0.57
B. 1? hp; 1800 rpm; 20° BTDC
6-6 15.1 1.3 0.53 2.20 3.50
6-5 16.4 2.8 0.37 1.24 2.86
*6-7 18.1 4.2 0.32 1.29 1.50
C. 15 hp; 1800 rpm; 10° BTDC
6-9 13.4 0.4 0.81 16.80 2.04
6-10 14.4 0.6 0.40 6.23 2.70
*6-ll 15.7 1.7 0.25 1.42 2.70
*Best runs (tests resulting in lowest pollutant mass rates under the
range of utilized conditions).
221
-------
air-fuel ratio (on either side of the optimum value) can cause severe
changes in the HC and NOX mass rates.
Table 9.5 data demonstrate the effect of spark timing on pollutant
mass rates. Spark retard substantially reduces the hydrocarbon levels in
the exhaust at all power settings and for all air-fuel ratios. Its effect
on NOX is a function of air-fuel ratio and the engine manifold pressure.
If the power output of the engine is held constant by spark retarda-
tion and manifold pressure changes (part A), NO control at constant air-
fuel ratio is effective. If power is controlled by increasing the mixture
ratio (richer mixture) at constant manifold pressure, then spark retarda-
tion is also effective in NO control.
A significant effect of spark retard is the increase in exhaust gas
temperature with retardation. For the tests reported here, the tempera-
ture increased an average of 100°F per 10° of retard. Systems employing
catalytic converters where the conversion efficiency is severely tempera-
ture dependent will probably require some retardation to maintain con-
verter temperature.
Table 9.6 shows the dependence of emission mass rates on power output.
For the range of power settings investigated, NOX is most sensitive to
power output, increasing more rapidly with power than would be expected
on the basis of air flow rates. There appears to be a small increase in
HC flow with power; however, these variations may be caused by air-fuel
ratio effects which cannot be separated from the data. CO is very in-
sensitive to power level.
9.2.4 Conclusions
The test program demonstrated the magnitude of the "cold start
problem" of a spark ignition engine. For a conventionally carbureted and
choked engine, the CO and HC mass levels during the initial phases of
engine operation greatly exceed the 1975 standards. On the other hand,
mixture ratio control and spark timing can reduce or eliminate the problem
if there is no requirement of engine driveability, i.e., the engine can be
effectively uncoupled from the wheels during the warm-up period.
222
-------
u>
TABLE 9.5
EFFECT OF SPARK TIMING ON POLLUTANT MASS RATE AND EXHAUST TEMPERATURE
A. 15 hp; 1800 rpm; approx. constant air-fuel at 15.8
C.
Test Nos.
**
(6-5) (16.4)
(6-6) (15.1)
6-11
12 hp; 1800
3-5
3-8
11-1
9 hp; 1800
3-6
4-2
4-6
Air-Fuel
15.9
15.8
15.7
rpm; approx.
18
16.9
15.5
rpm; approx.
19.4
14.9
13.3
Spark
Degrees BTDC
30
20
10
Pollutant
HC
0.73
(0.37) n ., (1.
(0.53) U>^ (2.
0.25
Mass Rates
CO
1.25
24), 7? (2
Z0y'/d (3
1.42
(grams /mi n)
NOX*
4.54
.86) . lft
.50) J'ia
2.70
Exhaust
Temperature
1260°F
1350°F
1460°F
constant manifold pressure
30
20
10
constant mani
30
20
10
0.52
0.34
0.23
fold pressure
1.80
0.26
0.21
1.21
1.11
1.13
1.72
0.85
2.20
1.00
1.08
1.48
0.25
1.05
0.95
1 300°F
1320°F
1400°F
1250°F
1320°F
1430°F
*as N02
**average of tests 3-4 and 8-2
-------
TABLE 9.6
EFFECT OF POWER ON POLLUTANT MASS RATES
A. 1800 rpm; approx. 16.5 air-fuel; 20° BTDC
Test Nos
5-7
6-5
9-4
B. 1800 rpm;
9-1
9-3
5-1
Powei
hp
10.5
15
23
approx.
11
14
20.5
Throttle
" Air Flow
Air-Fuel scfm
16.4
16.4
16.8
17.0 air-fuel;
16.6
17.5
17.1
28
30.2
44.8
20° BTDC
28.4
32.4
44.8
Pi
HC
0.16
0.37
0.59
0.49
0.40
0.94
Pollutant Mass Rates
(grams/min)
CO NOx*
1.06 0.27
1.24 2.86
1.69 5.72
*as NO,
1.10 2.00
1.21 1.98
1.87 5.30
.224
-------
Both cold start and hot engine pollutant mass rates are extremely
sensitive to air-fuel ratio:
• Rapid fuel mixture leaning to above stoichiometric air-fuel
ratios during cold start and reduce hydrocarbon and NOX levels
by more than 50%.
• Hot engine operation should be confined to lean fuel mixtures.
The optimum air-fuel ratio increases with power level settings.
With the spark at 30° BTDC and an engine speed of 1800 rpm, the
optimum air-fuel ratio lies between 16 and 17.5, depending on
the power setting.
Timing is important to emission control. During the cold period:
t Spark retard hinters smooth engine start-up and causes an increase
in the hydrocarbon emissions at any given power setting. It sub-
stantially reduces the NOX emissions during this period.
• Spark retard initiated approximately 30 seconds after engine
ignition appears to be the best timing schedule as it eliminates
difficulties associated with engine start-up, yet introduces NOX
control early and increases the exhaust gas temperature more
rapidly.
During hot engine operation:
• Spark retard at constant speed and power reduces the hydrocarbon
mass rate in the exhaust at all air-fuel ratios. NOX mass rate
is reduced by spark retard at constant power and air-fuel ratio.
CO is relatively unaffected by spark retard.
• Spark retard at constant speed, air-fuel ratio, and manifold pres-
sure substantially reduces HC and NOX exhaust concentrations;
however, engine power is reduced by approximately 33% per 10° of
retard.
Emissions are affected by engine power levels in the following: In-
creased engine power causes an increase in the emission mass rates. The
increase in exhaust NOX is disproportionately large with power increase.
Hydrocarbon emissions during the cold period are relatively insensitive
to power level.
Figure 9.7 summarizes the results of this study in terms of predicted
HC emission performance on a 1972 DHEW cycle. It is assumed that the
engine would be started cold and its choke, power and timing schedule
would be controlled during the cold period (three minutes) to produce
signatures as determined by the VW test results. For the remainder of
225
-------
EFFECT OF COLD START (3 MINUTES) HYDROCARBON
EMISSIONS ON TOTAL HYDROCARBONS EMITTED DURING
22 MINUTES (APPROX. 7 MILES) OF CONSTANT POWER
OPERATION OF A 1600 VW ENGINE
ro
ro
a*
10.5 HP;
15.0 HP;
23.0 HP;
22 MIN TOTALS
6.27 GRMS
12.43 GRMS
15.45 GRMS
10 15
TIME
20
25 MINUTES
COLD START PERIOD
FIGURE 9.7
-------
the cycle the engine would be held at power settings as shown in the
figure.
The total hydrocarbon emissions fail to meet the Federal standards
due to the cold start. Approximately 50% of the total emissions are
released by the end of the third minute of operation. However, during
the same period, a conventional engine would have emitted nearly twice
as much hydrocarbons and anywhere from three to ten times more CO.
Exhaust emissions can be minimized in the absence of exhaust control
devices by adopting the following procedures:
t Start the engine with a fully advanced spark at throttle positions
corresponding to 10-15 hp when the engine is hot.
• Retard the spark after 30 seconds.
t Relieve the engine choke rapidly during the first minute of opera-
tion while simultaneously increasing the throttle to the desired
final power.
t When the engine is hot, use a lean air-fuel ratio (16-17.5) or a
ratio closer to stoichiometry (^15.5) but with a spark retarded
10° or more from the normal operating point.
227
-------
9.3 Effectiveness of a Three-Component Catalyst on Exhaust Emissions
A three-component catalyst system was tested to determine its effective-
ness on a constant throttle, constant power engine. The three component
catalyst chosen was one of several produced by Universal Oil Products (UOP)
of Des Plaines, Illinois. The catalyst is reported to contain only non-noble
metals and for a narrow range of air-fuel ratios near stoichiometry is
claimed to have CO, HC and NO conversion efficiencies in excess of 90%.
TRW Systems personnel traveled to UOP to coordinate the testing. A
late model VW 1600 equipped with fuel injection and a UOP catalytic con-
verter/muffler was tested on a chassis dynamometer. Four hot runs were
made.
9.3.1 Hot Test Results
Table 9.7 summarizes the four tests. The first two runs were made at
very low power settings. The air-fuel ratio was apparently too lean for
the catalytic converter to reduce the NOX; i.e., there was insufficient
HC and CO in the exhaust stream. The low values of HC and CO are primarily
a result of lean engine operation and any catalytic conversion which took place,
In runs 3 and 4, the throttle was opened to a level more consistent
with EMT operation. At this setting, the car ran in second gear (auto-
matic transmission) between 26 and 28 mph and the chassis dynamometer
load was set at 6.2 to 6.4 hp. In run 3, the mixture ratio was again too
lean and the NOX is not controlled. A change in the air-fuel ratio
of about 6% to a richer mixture brought the system to a desirable state of
run 4. Contrasting the data of run 4 with hot test 5-6 of Table 9.4 in-
dicates the gain achievable with a catalytic device; a 25% reduction in HC,
a 20% reduction in CO and a 37% reduction in NOX.
9.4 Prediction of Emission Signature of an EMT System with a Three-
Component Catalytic Device
Table 9.8 presents predicted emission levels of an ICE-powered EMT
hybrid using a three-component catalytic converter of the UOP type. Pre-
dictions are based on combining the cold start test data, C-10, and run 4
of Table 9.7. The system fails to meet the 1975 standards for HC and NOX.
The primary source appears to be the three minute contributions of HC and
228
-------
TABLE 9.7
HOT TEST RESULTS OF UOP THREE-COMPONENT CATALYST ON VW 1600 ENGINE
Exhaust Emissions
grams/min
CO NOx
*JO 0.70
•tf 0.65
0.08 0.98
0.08 0.073
Run No.
1
2
3
4
Engine RPM
1800
1800
1800
1800
Manifold
Pressure
"Hg
12.3
11.8
8.25
8.5
Exl
HC
0.085
0.084
0.09
0.123
TABLE 9.8
PREDICTED EMISSION PERFORMANCE OF AN EMT HYBRID
ON A 1385 SECOND DHEW CYCLE*
(VW 1600 + UOP Catalyst)
HC CO N°x(aS N
grams/mile grams/mile grams/mile
Total Cycle 0.720 4.09 0.510
Percent Due to 58 95.5 66.5
Cold Start
•'engine operating at 14 shaft horsepower
229
-------
NO . In terms of the rapid rise in exhaust gas temperature and the delay
of NO appearance, it is probable that there would be some significant
A
catalytic conversion of N0tf during the three-minute cold period if the
^
catalytic device had been installed on the EMT equipment.
230
-------
REFERENCES
1. Gelb, Richardson, Wang, and Berman, "An Electromechanical Transmission
for Hybrid Vehicle Power Trains — Design and Dynamometer Testing",
Society of Automotive Engineers Paper No. 710235, January 1971.
2. Richardson, Gelb, Wang, and Licari, "System Design Implications of
Electric and Hybrid Vehicles", Intersociety Energy Conversion and
Engineering Conference, Denver, Colorado, August 1968.
3. Foote and Hough, "An Experimental Battery Powered Ford Cortina Estate
Car", Soceity of Automotive Engineers International Automotive
Engineering Congress and Exposition, Detroit, Michigan, January 1970.
4. Gelb, Richardson, Wang, and DeWolf, "Design and Performance Character-
istics of a Hybrid Vehicle Power Train", Society of Automotive Engine-
ers, Paper No. 690169, January 1969.
5. Byrne and Lacy, "Compatible Controller--Motor System for Battery
Electric Vehicles", Proceedings of the IEE. Vol. 117, No. 2, February
1970.
6. Kawai and Miyake, "Pulsating Current Traction Motors for the Electric
Multiple-Unit Coaches for the New Tokaido Line, Hitachi Review. Special
Issue on the New Tokaido Line, Tokyo, Japan, June 1964.
Further reading on chopper circuits for control of DC traction systems
(Section 5.1) can be found in:
Berman, "Battery Powered Regenerative SCR Drive", IEEE 1970 IGA Con-
ference Record, Chicago, Illinois.
Silicon Controlled Rectifier Designer's Handbook, R. Murray, editor,
Westinghouse Electric Corporation, April 1964.
SCR Manual. 4th Edition, F. Gutzwiler, editor, General Electric
Corporation, 1967.
Koch, "The Electric Car: A Design Challenge", Electro-Technology,
May 1968.
Morgan, "Time Ratio Control with Combined SCR and SR Commutation, IEEE
Transactions on Communications and Electronics, July 1964.
Heumann, "Pulse Control of DC and AC Motors with SCRs", Proceedings of
INTERMAG Conference. Paper No. 63-1011, April 1965.
231
-------
APPENDIX A-l
COMPARISON OF BOOST-BUCK AND BOOST PCUs FOR REGENERATION
Boost-Buck PCU
Figure 2.15 showed the simple circuit configuration for boost-buck
PCU. The power output of the generator per cycle must equal the power
delivered to the battery per cycle. Thus,
'on <* =VB '2 'off "*
where:
V = generator voltage
I, = generator current
VD = battery voltage
D
I- = battery current
t = time S, is closed
on l
t f- = time S, is open
T = t + t ff = chopper period
The current flows through the inductor continuously, whether or
not S, is open. In addition, if the inductor L is very large so there
is a very small ripple component, then
1^=12= constant (2)
and Vg 'l 'or ' VB 'l 'off
or
*"" - / (3)
VB
232
-------
The generator average current, I , is related to the time dependent
currents as follows:
t
=, on
ave "T" (4)
but, T
hence,
? > 'g
V
The peak current will be larger than the average by (1 + rp- ). The
B
rms current in the generator can be shown to be I = '.
thus
'g
V
Assuming the generator speed varies over a 10 to 1 range (as in the
case for the Speeder, slightly higher for the Torquer acting as a
generator), V can go from V , the generator voltage at lowest speed*
to 10 VQ. In addition, 1t is reasonable that in a boost-buck system
the battery voltage would be half the maximum generator voltage so
that
VB = 1/2 10 VQ
Using equation (6)
At 10% mas rpm Ig rm$ = Ig aveVl + 0.2 - 1.09 Ig
At 50% max rpm Ig rms = Ig aveVl + 1 - 1.414 Ig flve
At 100% max rpm I = Ig ayel 2 = 1.732 Ig aye
Since the ohmic losses in the generator are roughly proportional to the
square of I , the copper dissipation in the machine at top speed
will be three times larger than that due to the average current.
233
-------
The inductor conducts current I, at all times. From equation (5),
it can be seen that at top speed the inductor carries
<1 • 'g ave (1 + VJ[ > ' 3 ', ave-
and the inductor will have to be designed for three times the rated
machine current to prevent saturation.
The power delivered by the generator to the battery per cycle is
p = V I off = _PJ1 » I (7)
HB VB h T T Vg M (/)
or
rms
ft
- T J _?j
Mlf T
or ZB rms
thus
"V
rms B
234
B " V (8)
B
substituting from equation (5)
Pn = I V
B g ave g
The average current in the battery is:
, . , ittt. , in .!a. , !a
B ave IT 1 T Vn g ave VR
The rms current in the battery is:
-------
Boost PCU
Using the assumptions and nomenclature of the previous section
V I = v I off
Vg *1 VB M T
or
>g ave
g rms g ave ^ '
Since the inductor always carrier current, the inductor will be alway:
rated for the same average current as the generator.
The average battery current is simply
off = T -9-
ave g ave
and the rms battery current
rms
235
-------
APPENDIX A-2
MATHEMATICAL MODEL OF TORQUER SYSTEM
To facilitate analysis, it is convenient to divide the total chopper
period, T, into several subintervals. The current flow in each time
interval is determined as follows:
(1) When the power SCR, is closed and the commutating capacitor is
fully charged (t>0), the only circuit which contains electric current is
shown in Figure A-2.1. Therefore,
where ^ = R, + R., is total resistance of the series circuit, Rn the
resistance of the battery, and RM the resistance of the series motor.
During the time interval of interest, assuming the machine is saturated,
and the machine speed is constant, then E ^ constant, the solution of
the equation becomes
V - E - -- t
In most cases IM is negative and its value is determined by the
boundary conditions to be discussed later.
(2) At t = t-|, the commutating SCR 1s turned on and Cc dis-
charges through the linear Inductor L . S , the saturable inductor is
C A
blocking (high impedance) at this instant because it has been driven
out of saturation. The circuit diagram is shown in Figure A-2. 2.
When t>tj, the current iwt\ continues to flow in the battery-
armature circuit and an additional current i will be started in th
commutating circuit. This current is determined by the equation:
di
- Vb = 0 (3)
236
-------
SCR
ro
CO
— VB
-o
i ~0
c
FIGURE A-2.1
-------
ro
GJ
oo
"B
vw
SCR]
-O 5-
-------
Where V. is the initial voltage across the capacitor at t = t-j. By e
change of variable t1 = t - t, the initial conditions are:
i) i =0 when t1 = 0
C (4)
d1c
ii ) L r + 0 + 0 - V = 0 when t1 = 0
c
di
c
V
b
(5)
t'=o UG
Differentiating equation (3)
di ,
. _ T ."'it1 mot1
therefore 'c " M e + \2 e
"Rc ±VRc2 - 4Lc/Cc
where m, , = — %•. -—— = a, +
'•* ZLc ' ~
Substituting (4) and (5) into (7), we have
VB Olt'
ic = jjp- e sinWlt'
(note: o^ is negative in equation (8) )
Because of the diode in the commutating circuit, i cannot be
negative. The current i stops flowing at t9=ir/w, ^ m/L C , if R is
O c I ^~ ^ C C
small. At the moment current stops commutating SCR is turned off. A
reversed voltage VG (the voltage across the capacitor) is applied to
the power SCR^ to commutate it off, and the circuit goes into a third
mode of operation.
The voltage across the commutating capacitor C during the commu-
tation period is
ec -rc f fc(t)dt - \
JQ
239
-------
S1n
cos
wit)tw' -vb]|
(9)
when
t' = 0, i.e., t=t^, ec=-Vb; when t' = t^, i.e., t=t1 + t,,;
(10)
Furthermore if
a, « w.
-c(f=tc) = —
^L
-------
If
If
/°r
»
(14)
The current in the armature circuit at t = t-, -*• t^ is
Ra
(3) When t ^Ctj+tg), both SCRc and SCR1 are turned off, the power
circuit becomes that shown in Figure A-2.3. The saturable reactor S
J\
at this time is driven into saturation and its impedance becomes
approximately zero. The capacitor continues to discharge through the
motor. Therefore,
VV V2 + La^f + rj[+t '2dt-vo (16)
c tj+tg
Where V is the initial voltage across the commutating capacitor at
t=t2+tT By a chan9e °f triable t" = t (tp+t,), and differentiating
equation (16), we have
The initial constants are as follows:
241
-------
SCR
M
^/w•
ro
4k
ro
SCR
O
*
FIGURE A-2.3
-------
*ne va^ue given in equation (15)
di
t,,=0
assuming V^ ^ V. and
0 D
"a 'o « 2vb - Eg
dt77
t"=0 'a
i r (2Vb-
The solution of (17) is
") = !3e
n2t"
(18)
where
-R
4L
M.2
2L
Substituting the initial conditions into equation (18) we have
= 'oe
!t" (- ?
\ WJ
sin Wt" + cos
V")
(19)
2V E
Mb ,
W2 La
sinw,t"
2
or
(20)
243
-------
If during the time interval of interest
o9t cu 2V. - En
e ^ « 1, and -f- I « —2-,—9.
w« o W0 L
& c a
2V. - E (21)
sinw2t"
(4) The free wheeling diode will start to conduct when
La dP" * Eg - " ] volt -°
Using the approximate expression of i'2 given in equation (21),
equation (22) becomes:
(2V. - E ) cos w,t" - w. L I sin w,t" = - E (23)
D y L. caO c. y
Solving equation (23) for t" - t_, the new circuit becomes that shown
in Figure A-2.4.
The current which continues to flow in the armature circuit when
the diode conducts is determined by the equation
E + L + R1 = ° <24)
and its solution is
R
M
n
(25)
where
(t}+ t2+ t3) < t" < T
244
-------
ro
*>
en
"B
vs/v-
— VB
^c.
M
-wv
FI3U=E A-2.
-------
The constants IM in equation (25) and IM in equation (2) are determined
by the following conditions:
t3)
Current will continue to flow from the battery until the capacitor
C is fully charged. The battery current is determined by the equation
• RB 14 + t <4 dt + Vc (26)
where V is the initial voltage of the capacitor when t = t,+ t«+ t_
i.e. ,
1 rW^
c(t1+ t2+ t3) = C^ J 1
t? t2
dt • Vb (27)
The solution of equation (26) is
(28)
» . u I
where
(tj + t2+ t3) < t < T
Since Rg C ^ 0 this current decreases to zero very rapidly. The
overall current waveforms are shown in Figure A-2.5.
246
-------
CURRENT IN SCRc/ ?c (t)
ii
ro
ro m
-J Z>
i
rvj
I
Jl
BATTERY CURRENT, ig( t)
K>
If
00""
L
MOTOR CURRENT, i (t )
c
SCRc ON
SCR,, SCRc OFF
FREE WHELING
DIODE ON
-------
The equations
12(t) - - IQ cos w2f t - IQ + ll sm w2t" e
t"
2V- E
sm w2t" e
(20)
RM
- t"
IM e La (25)
M3
(t^ if tj) < t" < T
were programmed on the TRW-UCSB on-Hne computer, t-j is preselected
by the operator, t2 is computed from the natural frequency of the
commutating circuit (tg *_ ir AC C ) and t, determined from the equation:
(2Vb- Eg) cos w2t" - w2 La IQ sin w2t» = - Eg (23)
The constants IM , I and IM are determined by the conditions that the
current must be continuous 1n an inductive circuit and that ii(0\ =
248
------- |