EPA-460/3-76-0(M-a
February 1976
DEVELOPMENT
OF SPECIFICATIONS
FOR A MOTORCYCLE
DYNAMOMETER
AND MOTORCYCLE
COOLING SYSTEM:
VOLUME I- DESIGN STUDY
I .S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Mobile Souree Air Pollution Control
Emission Control Teelmolo^y Division
Ann Arbor, Miehigan 18105
-------
EPA-460/3-76-004-a
DEVELOPMENT
OF SPECIFICATIONS
FOR A MOTORCYCLE
DYNAMOMETER
AND MOTORCYCLE
COOLING SYSTEM:
VOLUME I- DESIGN STUDY
Robert J. llcrling
Olson laboratories. Inc.
121 Knsl Cerrilos Avenue
Anaheim. California *)2HO.">
Contract No. 6K-03.2 I II
Program KU'iiu-iH Mo. 2-AH-I30
EPA Project Officer: Glenn Thompson
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Offiee of Air and Waste Management
Offiee of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 48105
February 1976
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - as supplies permit - from the
Air Pollution Technical Information Center, Environmental Protection
Agency, Research Triangle Park, North Carolina 27711; or, for a fee,
from the National Technical Information Service, 5285 Port Royal Road,
Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Olson Laboratories, Inc., Anaheim, California 92805, in fulfillment
of Contract No. 68-03-2141. The contents of this report are reproduced
herein as received from Olson Laboratories, Inc. The opinions, findings,
and conclusions expressed are those of the author and not necessarily
those of the Environmental Protection Agency. Mention of company or
product names is not to be considered as an endorsement by the Environmen-
tal Protection Agency.
Publication No. EPA-460/3-76-004-a
ii
-------
PREFACE
This dynamometer development report is a compilation
of the task report submitted to the EPA by Olson Laboratories,
Inc., in partial fulfillment of Contract No. 68-03-2141.
The report is being published in the present to
expedite dissemination of the information. However, since
this form does not incorporate current EPA opinion or possible
recent modifications, the opinions, conclusions, and recom-
mendations of each chapter should only be considered as
those of the contractor when the task report was written.
This development report is intended to compile the
data which, with the addition of EPA analysis, formed the
basis for the EPA Dynamometer Specification. As an example
of the EPA analysis, the EPA has determined that motoring
ability of the dynamometer would greatly facilitate the
calibration procedure, and is, therefore, desirable for the
EPA dynamometer. The addition of this requirement may
significantly alter the relative ranking of the dynamometer
systems considered in Task 7. The Dynamometer Specification
should be considred as a reference for the current views of
the EPA motorcycle dynamometer requirements.
111
-------
TABLE OF CONTENTS
PREFACE iii
Section
1 TASK 1 - REVIEW OF REQUIREMENTS 1-1
1.1 Introduction 1-1
1.2 Task 1 - Report Objectives 1-1
1.3 Review of Requirements 1-2
1.3.1 Definition 1-2
1.3.2 Test Cell Environment 1-2
1.3.3 Emission Test Equipment 1-3
1.3.4 Motorcycle Dynamometer 1-3
1.3.5 Engine Cooling System 1-4
1.3.6 Driving Cycle for Emission Test 1-5
1.3.7 Driving Cycle for Mileage Accumulation . . 1-5
1.4 Safety Considerations 1-6
1.5 System Adaptability for Mileage
Accumulation Tests 1-8
1.6 System Adaptability for Testing
Three-Wheeled Motorcycles 1-8
2 TASK 2 - EVALUATION OF DYNAMOMETER POWER
ABSORBERS 2-1
2.1 Introduction 2-1
2.2 Review of Requirements 2-1
2.3 Terminology 2-2
2.4 Load Determinations 2-2
2.5 General Specifications 2-5
2.5.1 Power Capacity 2-10
2.5.2 Roller Diameter 2-11
2.5.3 PAU Controls 2-12
2.6 Power Absorption Unit Considerations .... 2-12
2.6.1 Utilities Required . . . 2-13
2.6.2 Moment of Inertia 2-13
2.6.3 Thermal Dissipation Effects 2-14
2.6.4 Inertia Simulation 2-14
2.6.5 Calibration 2-14
2.6.6 Calibration Stability 2-15
2.6.7 Maintenance. 2-16
2.6.8 System Adaptability 2-16
2.7 Characteristics of Candidate PAU's 2-16
2.7.1 Air Blower 2-20
2.7.2 Friction Brake 2-22
2.7.3 Hydraulic Pump 2-25
2.7.4 Hydrokinetic - Open Flow 2-27
2.7.5 Hydrokinetic - Closed Flow 2-29
2.7.6 Eddy Current 2-35
2.7.7 DC Motor/Generator 2-37
2.7.8 AC Adjustable-Speed Drives 2-38
2.8 Power Absorption Systems - Rankings 2-41
2.8.1 . Ranking Alternatives - Flywheel Inertia
Simulation 2-41
iv
-------
2.8.2 Ranking Alternatives - Partial PAU
Simulation of Inertia 2-42
2.8.3 Discussion 2-43
3 TASK 3 - EVALUATION OF ROLL CONFIGURATIONS . 3-1
3.1 Introduction 3-1
3.2 Review of Requirements 3-1
3.3 Quantification of Tire/Roll Losses 3-2
3.3.1 Test Description 3-3
3.3,2 Roller Test Results 3-4
3.4 Roller Surface Contour 3-13
3.5 Surface Finish 3-15
3.6 Material of Construction - Roller 3-16
3.7 Roller Mounting 3-17
3.8 Recommendations 3-17
4 TASK 4 - EVALUATION OF INERTIA SIMULATION
METHODS 4-1
4.1 Introduction 4-1
4.2 Review of Requirements 4-1
4.3 General Specifications 4-2
4.4 Inertia Simulation System Characteristics. . 4-4
4.4.1 Mechanical Inertia Simulation System . . . 4-4
4.4.2 PAU Inertia Simulation 4-7
4.4.3 PAU-Type Inertia Simulation with Added
Flywheel Inertia 4-9
4.4.4 Costs 4-9
4.5 Recommendations 4-10
4.5.1 Ranking Alternatives 4-10
4.5.2 Discussion 4-10
5 TASK 5 - EVALUATION OF VARIABLE FLOW
COOLING SYSTEM 5-1
5.1 Introduction 5-1
5.2 Review of Requirements 5-1
5.3 General Specifications 5-2
5.4 Description of Existing Dynamometer/Blower
Systems 5-2
5.5 Blower Characteristics 5-5
5.5.1 Terminology 5-6
5.5.2 Centrifugal Blowers 5-7
5.5.3 Axial Blowers 5-19
5.5.4 Mixed Axial Centrifugal Blowers 5-27
5.6 Variable Flow Control Method
Characteristics 5-27
5.6.1 Control Methods 5-29
5.6.2 Control Mechanisms 5-32
5.7 Recommendations 5-52
5.7.1 Ranking of Alternatives - Blowers 5-52
5.7.2 Ranking of Alternatives Variable Flow
Control Methods 5-53
5.7.3 System Recommendations 5-54
6 TASK 6 - COOLING SYSTEM ANALYSIS 6-1
6.1 Introduction 6-1
-------
6.2 Review of Requirements 6-1
6.3 Review of Task 5 Report Recommendations. . . 6-2
6.4 Design Alternatives 6-4
6.4.1 Ranking System 6-5
6.5 Evaluation Criteria - Blower 6-5
6.5.1 Flow as a Function of RPM and AP 6-5
6.5.2 Fan Wheel Rotational Inertia 6-8
6.5.3 Delivery Schedule 6-12
6.5.4 Noise 6-12
6.5.5 Reliability/Maintenance 6-14
6.5.6 Adaptability 6-15
6.5.7 Power Consumption/Efficiency 6-16
6.5.8 Motor Compatibility 6-21
6.5.9 Size and Weight 6-25
6.5.10 Flow Conditioning 6-26
6.5.11 Costs 6-27
6.6 Evaluation Criteria - Motor Control and
System Performance 6-27
6.6.1 Accuracy 6-27
6.6.2 Speed Range 6-29
6.6.3 Response Time 6-31
6.6.4 Delivery 6-33
6.6.5 Size and Weight 6-33
6.6.6 Adaptability 6-33
6.6.7 Reliability/Maintenance. 6-35
6.6.8 Interferences 6-36
6.6.9 Costs 6-36
6.7 Outlet Duct Design • 6-37
6.7.1 Outlet Configuration of Ducting 6-38
6.7.2 Velocity Profiles 6-41
6.8 Conclusions 6-41
7 TASK 7 - DYNAMOMETER SYSTEM ANALYSIS .... 7-1
7.1 Introduction 7-1
7.2 Review of Requirements 7-1
7.3 Review of Recommendations from
Tasks 2, 3, and 4 7-2
7.4 Design Alternatives 7-3
7.4.1 Evaluation Criteria and Ranking System . . 7-4
7.5 Evaluation of Dynamometer Systems 7-5
7.5.1 DC PAU - Interia Simulation by Partial
Mechanical Plus Electrical Techniques . . 7-5
7.5.2 DC System with Mechanical Inertia
Simulation 7-11
7.5.3 Eddy Current PAU with Flywheels 7-16
7.6 Conclusions and Recommendations 7-22
APPENDICES
A Motorcycle Test Summary A-l
VI
-------
LIST OF ILLUSTRATIONS
Figure Page
2-1 Motorcycle Road Load Force 2-4
2-2 Net Road Load Determinations (Honda 90). . . 2-8
2-3 Net Road Load Determinations (Harley
Davidson 1200 cc) 2-9
2-4 Basic Dynamometer Systems 2-17
2-5 Dynamometer Control System 2-19
2-6 Block Diagram of Control System 2-23
2-7 Hydrokinetic - Open Flow Control Systems . . 2-28
2-8 Power Absorption Unit - Cross Section View . 2-30
2-9 Principles of Dynamometer Operation 2-32
3-1 Honda Motorcycle Data Single Roll (Input
Torque) 3-6
3-2 Honda Motorcycle Data Single Roll (Wheel
Speed) 3-7
3-3 Harley Davidson Motorcycle Single Roll
(Input Torque) 3-9
3-4 Harley Davidson Motorcycle Single Roll
(Wheel Speed) 3-10
3-5 Honda Motorcycle Data 3-11
3-6 Harley Davidson Motorcycle Data 3-12
3-7 Roll Surface Contour 3-14
5-1 Centrifugal Blower Airfoil Impeller 5-12
5-2 Centrifugal Blower Backward Inclined/
Backward Curved Impeller 5-13
5-3 Centrifugal Blower Radial Impeller 5-14
5-4 Centrifugal Blower Forward Curved Impeller . 5-15
5-5 Axfal Blower Propeller 5-22
5-6 Axial Blower Tubeaxial 5-24
5-7 Axial Blower Vaneaxial 5-26
5-8 Mixed Axial Centrifugal Blower 5-28
5-9 Open-Ended Control System 5-31
5-10 Closed-Loop Control System 5-31
5-11 Parallel Blade Outlet Damper 5-35
5-12 Opposed Blade Outlet Damper 5-36
5-13 Inlet Damper 5-38
5-14 Comparison Curves of Inlet and Outlet
Dampers 5-39
5-15 Bypass Vane Control System 5-41
5-16 Eddy Current Drive 5-45
5-17 Basic Velocity Control Loop Eddy Current
Drive 5-47
5-18 Typical Interconnection Diagram 5-49
5-19 DC Regenerative Control System 5-51
6-1 Static Pressure vs. Fan Outlet Area 6-9
6-2 Fan Flow vs. Rotational Speed 6-10
6-3 Vaneaxial Blower Performance 6-17
6-4 Blower Power Requirements vs. Blower Type
and Size 6-19
6-5 Blower Efficiency vs. Blower Type and Size . 6-20
vn
-------
6-6 Centrifugal Blower Motor Mounting
Configurations 6-22
6-7 Controller and Operator Station Dimensions . 6-34
6-8 Air Discharge Pattern 6-39
6-9 Cooling System Layout 6-40
6-10 Qualitative Comparison of Laminar and
Turbulent Velocity Distributions 6-42
6-11 Velocity Distributions for Unsteady State
"Start-Up" Flow in a Circular Tube 6-43
7-1 Typical Torque Curve of Standard Dynamometer
(Eddy-Current Type) 7-18
7-2 Eddy Current Brake Cooling System 7-20
LIST OF TABLES
Table Page
2-1 EPA Motorcycle Performance Data 2-3
2-2 Motorcycle Road Load Requirements 2-6
2-3 Comparison of PAU Systems 2-21
2-4 Harley-Davidson - 1200 CC/Honda - 90 CC. . . 2-34
2-5 Price and Specification for Emission
Dynamometer 2-39
2-6 Price and Specification for Durability
Dynamometer 2-40
4-1 General Specification Flywheel Inertia
System 4-3
4-2 General Specification Electrical Inertia
Simulation. . . 4-3
4-3 Price and Specification for Emission
Dynamometer 4-11
4-4 Price and Specification for Durability
Dynamometer 4-12
5-1 General Specifications 5-3
5-2 Characteristics of Centrifugal Blowers . . . 5-8
5-3 Typical Sound Power Level Reductions by Fan
Silencers 5-16
5-4 Characteristics of Axial and Mixed Axial
Blowers 5-20
5-5 Summary of Control Systems 5-34
6-1 Task Report 5 - Recommendations 6-3
6-2 Unweighted Ratings 6-6
6-3 Fan Wheel Acceleration Rates (Zero to
Max RPM) 6-11
6-4 Typical Delivery Information 6-12
6-5 Noise Levels 6-13
6-6 Maximum Safe Wheel RPM of SWSI Blowers . . . 6-15
6-7 Motor/Vaneaxial Fan Compatibility 6-24
6-8 Motor Weights 6-26
6-9 Motor Specifications 6-29
6-10 Motor/Controller Costs 6-37
6-11 Weighted Ratings . 6-44
7-1 System Analysis Factors 7-6
7-2 Price and Specification for Emission
Dynamometer 7-12
7-3 Price and Specification for Durability
Dynamometer 7-13
viii
-------
Section 1
TASK 1 - REVIEW OF REQUIREMENTS
1.1 INTRODUCTION
The objective of this phase of the program is the
development of specifications for a motorcycle dynamometer
and motorcycle cooling system to be utilized in emission
certification programs. The specifications must address all
the attendant problems of dynamically simulating vehicle
road conditions. In the development of dynamometer specifica-
tions, various power absorbers, roll assemblies and inertia
assemblies will be evaluated and related to actual road-load
data. Variable-flow blower systems which simulate actual
engine cooling effects will be examined. Specific parameters
to be studied include blower style, ducting requirements,
noise levels, efficiency, power requirements, flow control
methods, cost, and delivery. In addition to establishing
the specifications for the dynamometer and the cooling
system, the general test cell layout, the requirements for
utilities, and the controls for the test cell's environment
will be examined.
1.2 TASK 1 - REPORT OBJECTIVES
This section will summarize the pertinent require-
ments and examine the effect these will have on the design
of the dynamometer, the cooling system and auxiliary systems
1-1
-------
Where possible, general and preliminary specifications will
be presented and general design criteria will be examined.
1.3 REVIEW OF REQUIREMENTS
The EPA's "Proposed Rulemaking (NPRM) for New
Motorcycles," dated 22 October 1975, and the system require-
ments, as outlined in the Request for Proposal and contract,
were reviewed. Relevant requirements are listed below, and
their impact on the design and specification for the motorcycle
dynamometer and cooling system are discussed.
1.3.1 Defini tion
A motorcycle is defined in the NPRM as any motor
vehicle designed to operate on not more than three wheels in
contact with the ground which is not a passenger car or
passenger car derivative. Additionally, for the purpose of
testing, the engine displacement must be 50 cc or larger,
and the cycle will be street legal. This definition specifi-
cally includes three-wheeled cycles. The implications of
designing a facility for testing for both two- and three-
wheeled motorcycles will be discussed in Section 1.6.
1.3.2 Test Gel 1 Environment
The NPRM specifies that all test phases be conducted
with an ambient temperature range between 20°C and 30°C.
These temperature limitations apply to both the test cell
itself and any adjoining space where the emission analysis
might be conducted.
No specific requirements are placed on humidity or
air contaminant levels within this facility; but the control
of these parameters is important. Reasonable humidity
1-2
-------
control will be achieved by adequately designing the heating/
airconditioning/ventilation system. Background air contami-
nant levels must also be maintained at low levels for reasons
of work safety and health and to provide reasonable back-
ground levels emissions analysis.
1.3.3 Emission Test Equipment
The scope of this program precludes the devel-
opment or specification of the emission test equipment.
However, because this equipment is to be placed in the
facility, space, utility and interface requirements for the
equipment must be considered. The following equipment may
be housed in the dynamometer cell and test facility:
• Constant Volume Sampler (CVS)
• Dilution air filter assembly
• Tail pipe connector and flexible tubing
• Exhaust gas analytical system housing nondis-
persive infrared analyzers for measurements
of carbon dioxide and carbon monoxide, a
chemiluminescent NO analyzer and an FID
A
hydrocarbon analyzer.
• Storage facility for 20 gas cylinders used in
the operation and calibration of the analyt-
ical system.
1.3.4 Motorcycle Dynamometer
The dynamometer requirements are discussed in
Section 85.478-15 of the NPRM and the Request for Proposal.
1-3
-------
These documents discuss the requirement for simulating
motorcycle inertia in 10 kg increments. The NPRM also calls
out a maximum inertial mass of 480 kg. However, as required
by the contract, overcapacity will be provided to a maximum
of 700 kg. The design of this assembly will be developed in
Task 4 of this report.
The NPRM does not specify the power absorption
unit or the relationship between load and vehicle speed. It
does define the road-load settings at 65 kph as a function
of vehicle mass and procedures for determining road-load
settings. The contract requires study and definition of the
power absorption system. This will be performed in Task 2.
The dynamometer roll configuration will be defined
in Task 3. However, the implications of testing three-
wheeled cycles will be discussed in Section 1.6 of this
report. The only other reference to the dynamometer in the
NPRM concerns the attitude of the motorcycle on the dyna-
mometer; and it is specified that the bike be level when
tested.
1.3.5 Engine Cooling System
The contract and the test procedure described in
the NPRM requires, a variable-speed blower system to simulate
engine cooling. According to the contract, at roll speeds
in excess of 10 kph, the duct exit velocity from the blower
should be within 10 percent of the corresponding roll speed,
and at roll speeds less than 10 kph, the air velocity
should be within ±1 kph. Also, the contract calls out an
outlet area of 0.5 m2. Additionally, the regulations specify
that the blower outlet will be between 0.15m and 0.2m above
floor level and that the blower, outlet will be squarely
positioned between 0.3m and 0.45m in front of the vehicle's
front wheel.
1-4
-------
The evaluation of blowers and fans for this appli-
cation will be presented in the review of Task 5 activities.
In this report, general specifications regarding the blower's
power requirements and positioning in the cell will be
presented. Additionally, the impact of utilizing the same
blower system for emissions tests of three-wheeled vehicles
and two-wheeled cycles will be examined.
1.3.6 Driving Cycle for Emission Test
Two driving cycles to be utilized in the emissions
test are defined. One is applicable to the smaller motor-
cycles, engine displacement less than 170 cc, and the other
to motorcycles with engine displacement 170 cc or greater.
The driving cycle for the larger vehicles requires equivalent
accelerations and decelerations, but higher maximum velocities
These velocities, accelerations, and decelerations will have
to be matched by the blower system for simulating engine
cooling. Maximum speed encountered during the test cycle is
90.1 kph (25.0 m/sec). The maximum sustained acceleration
occurs between 447 seconds of the cycle and 455 seconds when
the acceleration rate is 5.3 km/hr/sec (1.47 m/sec2). The
o
maximum deceleration rate is 5.3 km/hr/sec (1.47 m/sec ).
1.3.7 Driving Cycle for Mileage Accumulation
Mileage accumulation procedures are detailed in
the NPRM for motorcycles having an engine displacement less
than 170 cc and greater than 170 cc. Both the mileage
accumulation cycle and total amount of mileage accumulated
differs for the two engine classifications. The larger
motorcycles are subject to a more demanding test, requiring
an accumulation of 30,000 km at speeds reaching a maximum of
110 kph (30.5 m/sec).
1-5
-------
A dynamometer designed for mileage accumulation
test will require increased blower and power absorption
capacities. Additional safety precautions also must be
considered. All of these implications are discussed in
Section 1.5. In this report, only a dynamometer for emission
testing will be discussed in detail.
1.4 SAFETY CONSIDERATIONS
A review of OSHA regulations and the California
Industrial Safety Code was conducted to determine applicable
regulations. Few of the provisions of these regulations
apply to the dynamometer testing of motorcycles. Those that
do apply will be examined below. In addition, safety con-
siderations peculiar to the testing of motorcycles also must
be examined.
The following sections of the OSHA regulations
affect the design of the dynamometer facility:
• 1910.36 (b) (8) requires two exit paths from
the area.
• 1910.95 defines limits of noise exposure
without supplementary protection. Noise
exposure limits are listed below.
Duration per Day Sound Level
(hr) (dBA)
8.0 90
6.0 92
4.0 95
3.0 97
2.0 100
1.5 102
1.0 105
0.5 110
0.25 115
1-6
-------
New regulations under consideration may
reduce these exposure levels and require the
installation of noise monitors.
• 1910.101 describes procedures for the storage
of compressed gases. These procedures have
been developed by the Compressed Gas
Association.
• 1910.157 defines the number and type of fire
extinguishers required and their location.
Of the OSHA regulations, noise control will be the most
difficult to satisfy. Noise will emanate from the blower
and the motorcycle engine and drive train. Reverberation
reduction by the use of acoustical cell walls will be help-
ful and noise will be a parameter considered in the selec-
tion of the blower system. The use of ear protectors may be
mandatory. Eye protection also should be considered because
of the possibility of particles or small objects being
carried in the cooling airstream.
The possible consequences of mechanical failures
of the motorcycle in terms of operator injury must be
evaluated. Such things are rear wheel lockup, tire failure
or drive chain failure, and fuel tank failure should be
anticipated and appropriate operator safeguards be incor-
porated in the design of the test facility.
In the event of such failures, it is appropriate
to perform emergency shutdown operations automatically, such
as stopping the vehicle engine, stopping the cooling air,
shutting off the fuel and bringing the rolls to stop, all
without operator action. Emergency buttons also should be
provided by which the operator and/or test observer can
initiate emergency shutdown in case automatic detection
devices fail to perform or a malfunction occurs which is not
sensed by the automatic devices.
1-7
-------
1.5 SYSTEM ADAPTABILITY FOR MILEAGE ACCUMULATION
TESTS
As discussed in Section 1.3, the mileage accumu-
lation test procedure subjects the motorcycle to higher
velocities and accelerations than occur in the emissions
certification driving cycle. Therefore, greater loads and
operating ranges will be required if the dynamometer faci-
lity is to be utilized in both modes. Specifically, in the
mileage accumulation cycle, the dynamometer will have to
absorb twice the road-load experienced in the emissions
driving cycle. In order to simulate engine cooling condi-
tions during the mileage accumulation test, the linear
velocity at the exit of the blower duct must be 30.5 m/sec.,
20 percent greater than for the emissions' cycle and requir-
ing a 25 percent larger motor, in terms of horsepower.
While it is a simple matter to equip the dyna-
mometer and cooling system with components to meet the
requirements imposed by both test modes, the human safety
factor becomes significantly more critical during the
mileage accumulation test. Over 1,000 hours will be required
to accumulate 30,000 km, versus less than 1 hour to complete
an emissions test. It can be assumed that component failure
is more likely to occur in the process of mileage accumu-
lation than during emissions measurement. In addition,
driver fatigue will increase safety risks and the likelihood
of errors.
1.6 -SYSTEM ADAPTABILITY FOR TESTING THREE-WHEELED
MOTORCYCLES
The NPRM applies to three-wheeled motorcycles as
well as two-wheeled bikes. Therefore, general design para
meters which will be affected by the inclusion of these
vehicles must be examined.
1-8
-------
The design modifications necessary to test both
two- and three-wheeled motorcycles on a chassis dynamometer
are:
t Provision for added roller or rollers to
engage the two drive wheels that will span
the minimum and maximum tread dimensions.
• Realignment of cooling air discharge nozzle.
• Relocation of vehicle restraint system (e.g.,
front wheel clamp or cables).
0 Relocation of driver's aid and controls.
• Means for disconnecting added roller to
reduce windage, friction and inertia when
testing two-wheeled motorcycles.
• Increased test cell and pit width.
• More extensive protective covers or floor
plates.
In light of these design problems and the resul-
tant design complexities, this report will limit itself to
the examination of a suitable design for performing emis-
sions measurement on two-wheeled, motorcycles.
1-9
-------
Section 2
TASK 2 - EVALUATION OF DYNAMOMETER
POWER ABSORBERS
2.1 INTRODUCTION
This task report will present an analysis of power
absorption units (PAU's) and PAU control systems with respect
to their suitability for use in duplicating motorcycle road-
loads in performing exhaust emissions test procedures, as
defined in the Notice of Proposed Rulemaking (NPRM) and EPA
Contract No. 68-03-2141. These requirements will be examined
and a general specification for the PAU and control will be
presented. A detailed analysis of PAU's and controls is
presented in this task report, and recommendations are made.
2.2 REVIEW OF REQUIREMENTS
The NPRM defines the driving cycle for emissions
measurement in Appendix I and the durability driving schedule
in Appendix IV. The EPA project officer has indicated that
emissions measurements ajre of primary importance in the
design of the dynamometer, and the durability cycle is to be
considered as an added capability if found to be compatible
with the requirements.
Both procedures require the motorcycle to be
operated over a prescribed driving cycle which includes
idle, steady-state cruise, and acceleration and deceleration
2-1
-------
modes. The NPRM provides road horsepower loads for motor-
cycles according to classes of loaded vehicle mass at one
speed, 65 kph. This is analagous to the procedures employed
in the testing of light-duty vehicles. In automobile testing,
the usual practice has been to provide the inertia load
component by means of roller-driven flywheels and the road-
load component by means of a roller-driven PAU having a near
cubic/load characteristic.
2.3 TERMINOLOGY
The following terms and symbols are used in the
discussion of PAU's and control systems.
Power Absorption Unit (PAU): A device which
applies torque to the roller on which the motorcycle drive
wheel bears. (A PAU in combination with means for measurement
of torque and rpm is usually termed a dynamometer.)
Kilowatt (kw): Unit of power adopted by Systems
International (SI). One U.S. Horsepower (550 ft. Ibf/sec)
equals 0.7457 kw. One metric horsepower (HPm) equals 0.7355 kw.
2.4 LOAD DETERMINATIONS
As part of this contract, the EPA has supplied the
contractor with empirical road-load data for several motor-
cycles, expressed in the form of an equation which defines
the force required to move a motorcycle at any constant
t
speed. This equation and the empirical load factors for
several motorcycles are shown in Table 2-1. The relative
influence of each of these terms of the quadratic equation
is displayed in Figure 2-1 for three motorcycles.
Using the formulae summarized in Table 2-1, further
calculations were made to establish the road-load horsepower
2-2
-------
TABLE 2-1
EPA MOTORCYCLE PERFORMANCE DATA
(Received March 3, 1975)
F = F +
V + F2V
where V = meters/sec.
mass = kilograms
F = Newtons
NOTE: FQ to mass correlation coefficient is 0.92
Assume mass '= inertia; wheels, etc. neglected
MODEL
Suzuki 750
H.D.
Suzuki 50
Yam. Enduro
Honda CT 70
Yam 125 Rd
Kaw 175
Honda 360
345
395
159
195
135
205
195
245
57.3
54.3
23.3
21.2
24.3
28.2
24.9
26.0
0.0
2.02
.793
1.76
.609
1.50
1.98
2.51
0.412
0.362
0.281
0.300
0.234
0.303
0.276
0.265
SST/55(4)
2-3
-------
450 i-
400
350
300
co
§250
o
nc
O
200
150
100
50
(3) F = F0+FlV.+ F2V?
V = METERS/SECOND
(PLOTTED IN KPH)
HAR.DAV.-395KG
HONDA 360 - 245 KG
HONDACT70- 135 KG
I
I
20
40 60
KILOMETERS PER HOUR
80
100
120
Figure 2-1. Motorcycle Road Load Force
2-4
-------
levels as a function of speed for each motorcycle. These
data have been listed in Table 2-2. For each set of
horsepower/speed data a best fit power curve in the form of:
Y - kxc (2-1)
where Y » watts
x * speed (meters/second)
c » power coefficient
k » constant
was calculated for comparison with the existing load curve
for light-duty vehicle dynamometer, and this is also tabulated
in Table 2-2. From these data for the eight motorcycles, a
mean exponent was established and found to be 2.18 with a
standard deviation of 0.08 and a standard error of 0.29. As
a comparison, the mean exponent used to describe the perfor-
mance of the Clayton dynamometer used to test light-duty
vehicles is between 2.83 and 3.0. Results of tire-to-roller
losses, discussed in Section 3, combined with the road-load
horsepower predicted by the EPA empirical data are presented
in Figures 2-2 and 2-3. The predicted road-load, tire-to-
roller loss and net PAU horsepowers are plotted. These data
indicate that there is a considerable difference in the
exponential characteristic of the required PAU load between
the Honda 90 cc and Harley Davidson 1200 cc motorcycles.
2.5 GENERAL SPECIFICATIONS
Several factors have been considered in preparation
of general specifications for the PAU. These are:
t Power capacity required primarily for emissions
testing and secondly for durability driving.
2-5
-------
TABLE 2-2
MODEL
SUZUKI 750
HAR. DAV.
SUZUKI 50
YAM. ENDURO
MASS
Fl
M3TORCYCLE ROAD LOAD REQUIREMENTS
KILOMETERS PER HOUR
345 57.3 0.0 0.412
Y = (8,76 x. 10~4) x 2.08
395 54.3 2.02 0.362
Y = (1.11 x ID'3) x 2-04
159 23.3 .793 0.281
Y = (2.75 x 10-4) x 2-26
195 21.2 1.76 0.300
Y = (2.93 x 10-4) x 2-28
20
.529
70.0
0.0
12.7
57.3
76,
11.
11.
.579
.7
.2
,2
65.5
.275
36.4
4.4
8.7
27.7
.304
40.2
9.8
9.3
31.0
40
1.63
108.2
0.0
50.9
57.3
1.83
121.4
22.4
44.7
76.7
1.01
66.8
8.8
34.7
32.1
1.18
77.8
19.6
37.0
40.8
50
2.58
136.8
0.0
79.5
57.3
2.87
152.2
28.1
69.8
82.4
1.67
88.5
11.0
54.2
34.3
1.95
103.5
?4.4
57.9
45.o
65
4.70
191.6
0.0
134.3
57.3
5.13
208.8
36.5
118.0
90.8
3.17
129.2
14.3
91.6
37.6
3.70
150.8
31.8
97.8
53.0
85
9.21
287.0
0.0
22C.7
57.3
9.75
303.8
47.7
201.8
102.0
6.38
198.7
18.7
156.7
42.0
7.38
230.0
41.6
167.2
62.8
110
18.36
442.0
0.0
384.7
57.3
18.86
454.0
61.7
338.0
116.0
12.87
309.9
24.2
262.4
47.5
14.75
355.1
53.8
280.1
75.0
HP
F
FjV
F2V2
FO +
HP
F
FjV
F2V2
FO +
HP
F
FiV
FjV2
Fo *
HP
F
F,V
F?V2
Fo +
FjV
-------
IVJ
I
TABLE 2-2
MOTORCYCLE ROAD LOAD REQUIREMENTS (Continued)
KILOMETERS PER HOUR
MODEL MASS Fo FI F2 20 40 50 65 85 110
HONDA CT 70 135 24.3 .609 0.234
Y = (3.30 x 10~4) x 2-18
YAM. 125 RD 205 28.2 1.50 0.303
Y = C4.ll x ID'4) x 2-20
KAW 175 195 24.9 1.98 0.276
Y = (4.08 x KT4) x 2-20
HONDA 360 245 26.0 2.51 0.265
Y = (4.21 x ID'4) x 2-20
Y = (4.546 x 10-4) x 2-18 2 = .9641
x = 2.18 S = 0.0821 S x = 0.290
Mean Std. Deviation Std. error
.264
34.9
3.4
7.2
27.7
.347
45.9
8.3
9.4
36.5
.336
44.4
11.0
8.5
35.9
.341
45.1
13.9
8.2
36.9
.906
60.0
6.8
28.9
31.1
1.24
82.3
16.7
37.4
44.9
1.22
81.0
22.0
34.1
46.9
1.26
83.6
27.9
32.7
50.9
1.47
77.9
8.5
45.1
32.8
2.03
107.5
20.8
58.4
49.0
1.99
105.6
27.5
53.2
52.4
2.06
109.0
34.9
51.1
57.9
2.74
111.6
11.0
76.3
35.3
3.78
154.1
27.1
98.8
55.3
3.70
150.6
35.7
90.0
60.6
3.80
154.7
45.3
86.4
68.3
5.43
169.1
14.4
130.5
38.7
7.46
232.5
35.4
168.9
63.9
7.24
225.5
46.7
153.9
71.6
7.38
230.0
59.3
147.7
82.3
10.86
261.4
18.6
218.5
42.9
14.83
356.9
45.8
282.9
74.0
14.25
343.1
60.5
257.7
85.4
14.42
347.1
76.7
247.4
99.7
HP
F
FiV
F2V2 •
F + PiV
o I"
HP
F
FiV
F2V2
F0 + FiV
HP
F
FiV
F2V2
F0 + Flv
HP
F
FiV
F2V2
Ffj + F-iV
-------
HONDA 90
w
S
O
20
MOTORCYCLE HP
NET PAD HP= (Kph3-1:
TIRE/ROLL LOSS
40 60 80
SPEED - KILOMETERS/HOUR
100 120
Figure 2-2. Net Road Load Determinations
-------
HARLEY DAVIDSON 1200 CC
12
5
O
o:
UJ
?
O
a.
20
MOTORCYCLE
HORSEPOWER
NET PAU HP = (kph24 ;
TIRE/ROLL LOSS
40 60 80
SPEED - KILOMETERS/HOUR
100
120
Figure 2-a Not Road Load Determinations
2-9
-------
• Inertia simulation method.
• Roller diameter.
• Complexity of PALI controls required.
Many of these factors interact, which necessitate making at
least tentative engineering judgements to avoid an unmanage-
able selection matrix. These shall be discussed in the
following paragraphs.
2.5.1 Power Capacity
The required PAU capacity for road-load, as calcu-
lated from the EPA formula and data on eight motorcycles is
shown in Table 2-2. A maximum of about 14.7 kw (20 HPm) is
indicated, including operation at 110 kph. Therefore, a PAU
of 18.4 kw (25 HPm) absorption capacity would provide an
adequate reserve to perform both emissions and durability
cycles.
If the PAU is also to be used for inertia simulation,
it must be capable of both motoring and absorbing, and the
horsepower capacity must be increased according to the
amount of inertia simulation. Maximum PAU inertia simulation,
less the intrinsic inertia of the roller, PAU, etc., would
result in about 37 kw (50 HPm) absorbing and 18.4 kw (25 HPm)
motoring requirement if both emissions and durability cycles
were considered.
The increased cost of the PAU for inertia simulation
and the probable errors due to control and PAU response lag
as discussed in the Task 4 report, are considered sufficient
reasons to limit the PAU function to providing road-load.
2-10
-------
2.5.2 Roller Diameter
The roller diameter influences the PAU specification
because it determines the PAU rpm for a given equivalent
road speed, assuming that the PAU is directly coupled to the
roller shaft. The Task 3 report recommends a 530.5 mm
diameter roller, which results in 1,100 rpm at 110 kph.
Most PAU's of the required power capacity are
designed for operation at 4,000 to 6,000 rpm which could be
achieved by means of a gear box or other type of speed
increaser. However, it may not be possible to sufficiently
reduce the losses in such a speed increaser to the point
that the losses would be insignificant. Four PAU installa-
tions can be considered:
• Mount the gearbox on the nose of the PAU and
trunnion mount the combined unit.
• Foot-mount both the PAU and the gearbox and
connect the output shaft to the roller shaft
by means of a shaft torquemeter.
• Provide corrections to the PAU command signal
to compensate for predetermined gearbox
losses.
• Select a PAU of a larger rated capacity,
having sufficient torque load to absorb the
required power at the lower RPM.
The use of an integral PAU and gearbox does not
significantly degrade measurement accuracy. However, gearbox
losses do limit the minimum torque of a nonmotoring PAU and
necessitate operating a motoring PAU through the zero power
point. The use of a foot-mounted PAU and gearbox with an
2-11
-------
output shaft torquemeter yields similar results. Compensa-
tion of the PAU control for gearbox losses is poor because
the losses vary with temperature, rpm, and load. Therefore,
it is felt that the selection of an overrated PAU which has
sufficient torque load to absorb the required power at the
lower rpm should be pursued. Information on PAU's which can
absorb 18 kw at 1,100 rpm has been solicited from vendors
and is discussed in Section 2.7.
2.5.3 PAU Controls
The controls required to produce loading according
to an algorithm, such as that furnished by the EPA, must
have the capability to accept the load factors FQ, Fj and
F£ defined in Table 2-1. These are empirically determined
values representing a constant, a linear and a square term
in the formula for the road-load force for each individual
motorcycle. The control must then calculate the instantaneous
load force according to roller rpm and cause the PAU to
apply this force (torque) to the roller shaft.
The tire-to-roller loss will have to be determined
for multiple loads and tire tread constructions so that
accurate corrections can be applied to the load command.
2.6 POWER ABSORPTION UNIT CONSIDERATIONS
In this section parameters which affect the perform-
ance and selection of a PAU are considered. These include:
• Utilities required
t Moment of inertia of rotor
• Method of heat dissipation
• Simulation of vehicle inertia
2-12
-------
• Calibration
0 Stability of calibration
• Maintenance requirements
• Adaptability to test cycle revision
Detailed information on the performance of specific
PAU's is presented in Section 2.7.
2.6.1 Utilities Required
The discussion of utilities here is limited to
those for operation of the PAU and control.
Cooling water is generally required for removing
heat generated by power absorption in many of the PAU's
considered. Exceptions are the air brake, dry gap-type eddy
current absorber and DC motor/generator.
Water supply pressure of 200 to 350 KPa (30 to
50 psig) and volumetric rates, of 8.5 m /s per KW (0.1 gal/
min/HPm), are generally required. In any of these units a
small minimum flow of water is required even though no power
is being absorbed. The cooling water may be discharged into
the drain or it can be recovered and recirculated through a
cooling tower. The choice is made based upon local codes,
whether or not there is any existing cooling tower system
and/or ecological factors.
2.6.2 Moment of Inertia
The moment of inertia of the PAU rotor is noted
because it, in combination with the roller, shaft and coupl-
ings, constitutes the minimum equivalent vehicle weight
which can be tested with proper inertia simulation. Conver-
sely, it is of interest in determining the minimum flywheel
size for mechanical inertia simulation or the motoring and
absorbing power requirements for PAU simulation systems.
2-13
-------
2.6.3 Thermal Dissipation Effects
The dissipation of PAU heat by means of cooling
water has been discussed above. The Lear Siegler eddy
current PAU is cooled by radiation and by convection air.
This results in additional heat load on the system for the
control of ambient temperature in the test facility. DC
motor/generator type PAU's are usually regenerative systems
in which most of the power absorbed is returned to the
electric power lines. The heat developed in the armature
and field coils is dissipated by means of forced airflow
from self-contained blowers. Heat from this source is of
little consequence.
2.6.4 Inertia Simulation
Simulation of vehicle inertia is discussed in
Task 4. However, it is appropriate to indicate that if PAU
inertia simulation is used, then the PAU must be capable of
both motoring and absorbing modes of operation and the power
capacity must be increased by as much as 100 percent.
The DC motor/generator, universal eddy current and
hydraulic pump/motor types of PAU's, which can be used for
vehicle inertia simulation, all require, even for fixed load
control, control systems that are more complex than those
for similar operation of hydrokinetic PAU's. The additional
equipment necessary to perform PAU inertia simulation and/or
load control per the EPA algorithm would consist of a small
analog or digital computer to generate the command signals
for the controller.
2.6.5 Calibration
Calibration of the PAU system should involve all
the components in the chain, including the calibration arm,
2-14
-------
the PAU cradle bearings, the PAU to load cell linkage, the
load cell cable and connectors, the load cell power supply,
amplifiers, attenuator switches and readout instruments. As
an example, the torque measuring system should be checked by
the loading of certified weights on an arm extending from
the PAU and observing the display on the torquemeter.
Similarly, rpm should be checked by rotating the roller
shaft, illuminated by a line frequency strobe lamp, and
reading the rpm display. Calibration checks which employ
the shunting of a strain gauge bridge, application of refer-
ence voltage or frequency to an rpm indicator, etc., are
convenient and useful troubleshooting techniques and cali-
bration checks, but are unacceptable for basic calibration.
The above examples are, perhaps, obvious. However,
calibration of the mechanical or PAU-type inertia simulation
system should be given particular attention. The technique
and equipment for calibration of the system selected should
be included in the final design.
2.6.6 Calibration Stability
Stability of calibration of the elements of the
motorcycle dynamometer system is an important factor in the
selection of components. Short-term drift in zero, span or
linearity can be almost undetectable during a test run and
can seriously degrade test-to-test correlation. It is,
therefore, important to select system elements of known
stability and immunity from temperature, vibration, supply
voltage change, electrical "noise," etc. Special precautions
should be taken to avoid interferences from noises emanating
from high energy ignition systems.
2-15
-------
2.6.7 Maintenance
The frequency and amount of maintenance, aside
from calibration, determines the available operating time of
the facility and thus, the operating cost to a large extent.
This is believed to be particularly important when require-
ments for testing tend to be seasonal.
2.6.8 System Adaptability
The adaptability of the motorcycle dynamometer
test facility to changes in the test cycle and the vehicles
themselves is a valid design criteria. This influences the
choice of PAU and control primarily with respect to the
road-load characteristic. It is anticipated that fuel
supply, environmental and consumer protection considerations
will probably result in the need for increased precision of
duplicating actual road-loading for emissions and fuel
economy tests. Such requirements can be better met with a
PAU and controller capable of loading the vehicle per the
EPA algorithm in which the constant, linear and squared-load
factors can be adjusted independently.
Unfortunately, the PAU systems having this capa-
bility are the most costly and longest in delivery time.
The alternative, to improve delivery and reduce initial
cost, would be to select a PAU which can meet present EPA
requirements and can be upgraded by adding a more complex
control system.
2.7 CHARACTERISTICS OF CANDIDATE PAU's
A basic dynamometer system is-schematically shown
in Figure 2-4. This system consists of a direct-driven PAU,
a variable mechanical inertia system and a roller. The PAU
2-16
-------
VARIABLE
MECHANICAL
INERTIA
nni i CR
DA
rf\
1 1*
u
TAPH
o
TORQUE
SENSOR
(a)
oru i CQ
DA
rf\
1 1*
U
T Ar*ij
1 A(jH
FIXED
INERTIA
•POWER ABSORPTION UNIT
O
TORQUE
SENSOR
(b)
Fifur«2-4. IMIC DyiwmonMtw Systtim
2-17
-------
serves as the power absorber and the flywheel system simu-
lates vehicle inertia. A second dynamometer system is also
shown in Figure 2-4. It consists of a PAD, a roller, and a
fixed flywheel. In this second system, the PAU serves two
functions; power absorber and inertia simulator.
Some PAU's have fixed power absorption character-
isitics. PAU's of this type include fans, blowers, hydraulic
pumps and hydrokinetic units.
A more accurate duplication of actual road opera-
tion is possible by adding a control, such as that shown
schematically in Figure 2-5, which develops the speed/load
characteristic according to the EPA algorithm using the
empirical load factors FQ, F, and F^. The dynamometer
control system in Figure 2-5 merely shows the elements
required to operate the PAU on the EPA load algorithm. The
empirical load factors FQ, Fj and F2 are manually set for
the vehicle to be tested. Similar manual inputs are required
for the vehicle mass and the tire/roll loss characteristic.
Roller speed and torque signals are fed back and
compared with the program command. Errors or differences
cause the PAU torque to be changed in the appropriate amount
and sense. Other essential elements of the control system,
such as signal filtering, proportional band, derivative
function, etc., are not shown.
Inquiries regarding PAU characteristics were sent
to representative groups of manufacturers. While some of
the responses were limited, they were sufficient to permit
an examination of candidate units. Specific types of PAU's
described are:
• Air blower
t Friction brake
• Hydraulic pump
• Hydrokinetic - open flow
• Hydrokinetic - closed flow
2-18
-------
Fo Fi F2
VEHICLE
MASS
I
I
INERTIA
SIMULATION
RPM (kph)
TORQUE
TORQUE ERROR
LOAD COMMAND
TIRE/ROLL
CHARACTERISTICS
PAU
POWER SUPPLY/
CONTROLLER
Figure 2-5. Dynamometer Control System
2-19
-------
• Eddy current
• DC motor/generator
t AC adjustable speed drives
Their basic features are summarized in Table 2-3.
2.7.1 Air Blower
There are available several types of air blowers
which exhibit an approximately cubic speed/load character-
istic. Such a blower, driven by the roller upon which the
motorcycle rear wheel rides, would appear to provide a simple
low cost loading device for motorcycle testing. Unfortunatelyi
there are several problems which render this type of PAU
unsuitable for the accuracy and flexibility needed to perform
the contemplated tests.
The horsepower speed characterisite of an air
blower is essentially fixed. It would be necessary to
adjust the blower load to match the road-load characteristics
of the motorcycle to be tested. This could be accomplished
by throttling either the intake or discharge of the blower
if it were directly driven from the roller shaft. Alterna-
tively, an adjustable ratio drive could be interposed between
the roller and the blower. Neither solution is believed to
be practical. In addition, some type of damper control
would be required to alter the blower's inherent cubic
horsepower load curve in an effort to match the power curves
cited in Table 2-2.
Adjustable dampers or orifices can be used only
over a limited range without producing serious changes in
the characteristic speed load curve of the blower. Further,
they are susceptible to change in calibration with changes
in air density due to temperature and barometric pressure
variation.
2-20
-------
TABLE 2-3
COMPARISON OF PAU SYSTEMS
Utilities Required
Rotor
Inertia
Cooling
Air Brake
Typical
Hydraulic fume
Typical
Hydro* inetie-open flow
Go Power D-552
Industrial Dyno
Kahn Industries (1-124
N.A. (Large) Nom
Vehicle Inertia.
Simulation By PAU
No
Complex Speed/Load Calibration
Curve Stability
Required
Main*
No
Poor • Affected Low
By Air Density
IM
I
Cooling Water, Control
Electric Power
Cooling Water. Control
Electric Power
Coolinf Water. Control
Electric Power
Cooling Water. Control
Electric Power
N.A.
-------
An infinitely variable ratio drive of sufficient
power capacity would be costly. Changing the ratio between
the roller and the blower would also result in changing the
inertia of the system.
Because of the relatively low roller speed, the
size of a blower for direct drive may be excessive. A
speed increasing belt, chain or gear drive to permit use of
a smaller size blower would add cost, and modify the character-
istic load curve of the system because of the losses in the
drive. Compensation for changes and the other variables
would probably require a moderately complex control system,
which is not, to our knowledge, commercially available.
2.7.2 Friction Brake
The friction brake PALI is composed of two distinct
parts. The first of these is the rolling gear involving the
rolls, bearings, mechanical frame, which holds it all
together; and the mechanical parts of the brake discs and
brake pads. The second part is the control system which
electronically and pneumatically controls the application of
pressure to the pads in order to provide holding torque for
the vehicle under test. Figure 2-6 is a block diagram of
the control system. It can be seen that the control system
consists of two parts: first, an electronic part which
calculates the desired torque from the speed signal generated
by the dynamometer tachometer and then produces an electronic
signal corresponding to the error between the calculated and
measured torque; and secondly, a mechanical system which
converts this electronic error signal into an air pressure
command to the disc brake.
The mechanical system is composed of a solenoid
which is driven by the current output of the electrical
subsystem. The force on the plunger of this solenoid is
proportional to the current flowing through the solenoid.
2-22
-------
rsj
ro
co
ELEC/PRESSURE
TRANSDUCER
APPERATURE
SOLONOID
VALVE
AIR
SUPPLY
P SUPPLY
REGULATOR
FILTER
jTEST GAUGE TEST GAUGE
' POSITION A ADJ POSITION e
OFFSET
TORQUE
SPEED
O
(D
o
o
c
BRAKF.
CYLINDER
FORCE
TRANSDUCER
1
-Q
TACH
FIGURE 2-6. BLOCK DIAGRAM OF CONTROL SYSTEM
-------
However, the constant of proportionality between the mechani-
cal force and the current flowing depends on the position of
the plunger in the solenoid and is variable, to some extent,
from one unit to another.
The force of the plunger presses down on the top
of a small ball. The pressure below the ball builds up
until the air pressure on the bottom of the ball exceeds the
mechanical pressure from the solenoid on the top of the
ball. At this point, the ball rises off the seat and the
air leaks out of the system at a rate rapid enough to maintain
equilibrium. Thus, at all times, the mechanical pressure on
the ball from above is approximately equal to the air pressure
on the ball from the bottom. In order to compensate for
pressure differences from 0 to about 414 kPa (60 psi), the
ball needs to rise only a few thousands of an inch off the
seat in order to provide the correct leakage rate.
Since the air for the electrical-to-pressure
transducer is supplied through a very tiny orifice, the time
required to build up the pressure or decrease the pressure
in the electrical-to-pressure transducer would be very great
if it had to drive any substantial volume. For this reason
the output of the electrical-to-pressure transducer is
applied to the control input of an air relay. The air relay
has the ability to control a very large flow of air into the
pistons.
The output of the air relay drives the four pistons
on the brake assembly. The air relay is capable of both
increasing or decreasing the pressure corresponding to
changes in the inlet control pressure.
Intrinsically, friction brakes of the type described
result in linear load curves. However, Autoscan Inc., has
developed a Road Profile Simulator which allows the friction
brake to be programmed to match almost any shape of load
curve. An analog computer is used in conjunction with the
algorithm to generate a voltage signal which is proportional
2-24
-------
to torque. A sensor measures the torque being applied to
the brake and the two signals are compared, resulting in a
corrective error signal. Autoscan has produced units which
have selectable power curves. They have not developed a
unit which can accept algorithm constants, though this can
be developed.
One problem which has been experienced in previously
developed units is control loop instabilities. This princi-
pally occurs when the vehicle is subjected to prolonged and
pronounced accelerations and decelerations. In conversations
with Autoscan personnel they admitted that their unit had
not been used in conjunction with driving cycles and could
not match the performance and versatility of an eddy current
unit especially with respect to transient response
characteristics.
The cost of the friction brake is approximately
$3000-$4000. However, some development costs would be
incurred in the design of a versatile system which could
accept the EPA algorithm.
2.7.3 Hydraulic Pump
A hydraulic pump driven from the roller and discharg-
ing fluid through a variable orifice is a simple means of
providing vehicle load. This system is used in the Hartzell
Mark II motorcycle dynamometer and other commercially-
available tuneup dynamometers. With this unit, pump discharge
pressure and roll speed are used to obtain horsepower from a
nomograph. Horsepower could also be calculated automatically
from these data if the pump characteristics were known.
Better accuracy could be obtained using pump reaction torque
and roll speed directly. This basic system could be modified
by the addition of a shaft or pump reaction torquemeter and
an electrically actuated relief valve. With the addition of
the control in Figure 2-5, it could load the roller per the
EPA algorithm.
2-25
-------
In the event that a nonflywheel inertia system is
selected, the hydraulic system would become considerably
more complex. First, the pump would have to be of a type
capable of motoring, which complicates the control system.
Second, motoring would require the addition of a hydraulic
power supply.
Perhaps the best features of the hydraulic pump/motor
are its inherently low inertia and small physical size for a
given power capacity which make a high rate of response
achievable.
Disadvantages of the hydraulic dynamometer system
are:
• Potential safety hazard is somewhat higher
than that of hydrokinetic or electrical-type
systems due to presence of hot, high pressure
oil.
• Longer warm-up time is required to reach the
oil temperature and viscosity necessary for
proper response of control system.
t Additional controls are needed to ensure
stable temperatures within hydraulic system.
• The ancillary equipment such as the oil
reservoir, heat exchanger, control valves,
etc., require considerable space.
• Hydraulic systems tend to be sources of
noise.
t Sophisticated hydraulic systems, as described
above, have not been used before in this
application and may require extensive
development.
2-26
-------
2.7.4 Hydrokinetic - Open Flow
The term "hydrokinetic - open flow" is used to
describe power absorption units in which the working fluid,
usually water, is introduced between a set of rotating and
stationary discs within a circular housing. In these units
which exhibit a nearly cubic speed/load characteristics,
power is absorbed by the shear and turbulent flow of the
water between the discs. The discs are perforated to
increase the power absorbed in a given diameter and number
of discs. These PAU's are "open flow" systems. The fluid
is admitted into the housing near the axis of rotation and
it collects at the periphery of this housing by centrifugal
action. It is discharged at a controlled rate. The amount
of housing fill and, thereby, the amount of power absorption
is controlled by balancing the inlet and outlet flows.
Precise control of supply pressure is essential for accurate
and stable fill or load control. Water discharged from the
PAU may be collected in a sump, cooled, deaerated and recir-
culated to conserve water, or it may be discharged into a
drain. Schematics showing the operation of this type of PAU
used in conjunction with manual or automatic load control
systems are shown in Figure 2-7.
Hydrokinetic - open flow PAU's are relatively low
cost in terms of horsepower capacity when the load is control
led manually. Because of the "open flow" operation, load
index setting or polynomial speed/load operation require
somewhat elaborate closed loop control systems as shown in
Figure 2-7. The availability of control systems which can
accept the EPA algorithm terms is unknown. It is only
feasible to use hydrokinetic - open flow-type PAU's with
mechanical inertia simulation because they lack motoring
capability.
Go Power Systems furnished information on their
"52" series absorption units, and their Model 552 has suffi-
cient capacity for this application. Rotor inertia is
2-27
-------
MANUAL CONTROL SCHEMATIC
REF. C503-100-005
TRIM
CONTROL
VALVE
WATER SUPPLY
50PSI
ATMOSPHERIC
WATER DRAIN
INLET
CONTROL
VALVE
OUTLET
CONTROL
VALVE
WATER SUPPLY
50PSI
INLET CONTROL VALVE
MAGNETIC PICKUP &
60 TOOTH GEAR
DYNAMOMETER
OUTLET CONTROL VALVE
ATMOSPHERIC
WATER DRAIN
AUTOMATIC CONTROL SCHEMATIC
REF. C503 100-020
00
SPEED SET POINT
ELECTRONIC
CONTROLLER
Figure 2-7. Hydrokinetic-Open Flow Control Syttemt
2-28
-------
-2 2
3.95 x 10 kg.m . Cooling water flow required is approxi-
mately 1.58 x 10"5 m3/s at 37 kw (50 HPm). The price of the
basic unit is "in the $2,000 to $4,000 range."
Industrial Dynamometer Company submitted product
information, indicating that they would supply a PAU with a
capacity of 37 kw (50 HPm). Water flow would be 1.89 x 10"4
3 22
m /s and the rotor inertia is 1.171 x 10 kg .m . Response
time from zero to full load is stated to be approximately
two seconds. The price of this unit is $3,500.
Kahn Industries, Inc., quoted their Model 061-124
for the motorcycle PAU application. The rotor inertia of
this unit is 5.31 x 10"1 kg.m2. Water flow of 1.85 x 10"4
m /s/kw is required. Kahn also supplied information on
their manual and automatic load control systems. The price
of the Kahn PAU is $7,250 and the automatic set point load
control system is quoted at $4,877.
The hydrokinetic - open flow-type of PAU appears
less than optimum for this application because of the need
to develop a control system for maintaining a fixed exponen-
tial curve, other than cubic, or a complex speed/load
equati on.
2.7.5 Hydrokinetic - Closed Flow
The hydrokinetic closed flow type of PAU, exclusively
built by Clayton Manufacturing, differs from the hydraulic
friction type in that the working fluid moves spirally in a
vaned toroidal chamber which is divided into a static section
and a rotating section. The vanes are arranged so that the
fluid is accelerated by the rotating half of the chamber and
decelerated by the static half, thus developing reaction
torque. A cross sectional view of this type of PAU is shown
in Figure 2-8.
2-29
-------
PACK BUXtNGS AND MUFT
AS SHOWN USING SHELL
AI.VANU *1 CNCASC OH KUAL
WATEH HEAD
Figure 2-6. Power Absorption Unit Cross Section View
2-30
-------
Heat generated in the working fluid is dissipated
by means of a heat exchanger through which part of the
toroidal flow is shunted. Thus, the need to precisely match
open-flow hydraulic friction PAU's is eliminated. A schematic
showing the operational components of this system is shown
in Figure 2-9.
The torque developed at a given rotor rptn is
directly related to the mass of fluid in the torus and the
geometry of the vanes in both halves of the torus. The
speed/load curve characteristic is controlled in design and
manufacture to be slightly less than a cubic; actually 2.8
to 2.9.
Thus, for a given amount of fill, this type of PAU
will provide a specific and repeatable loading at each speed
within its operating range and the load will change with
speed according to the built-in power characteristic. If a
PAU having a power characteristic of 2.8 is "indexed" or
filled sufficiently to provide a load of 3 kw (4.1 HPm) at
65 kph, for example, it will provide the following loads at
other speeds:
SPEED LOAD LOAD
(kph) (kw) (HPm)
10 .02 .02
30 .34 .47
50 1.4 2.0
70 3.7 5.0
90 7.5 10.1
110 13.1 17.8
These speeds and loads will be duplicated until the index
point is changed by increasing or decreasing the amount of
fluid in the PAU.
2-31
-------
POWER
METER
COOLING WATER
SUPPLY
DRIVE ROLL
POWER ABSORPTION UNIT
(With Internal Heat Exchanger.)
STRAINER
FLOW CONTROL
VALVE UNLOAD
VALVE
LOAD WATER
DISPOSAL
Figure ^-9. Principles of Dynamometer Operation
2-32
-------
The loads exhibited by a hydrokinetic - closed
flow unit having an exponential characteristic of 2.83 has
been compared in Table 2-4 with the loads calculated from
the EPA generated algorithm. As shown in this table, the
errors in PAU load in the lower speed range are small in
horsepower, but large in percentage of the calculated PAU
load. The influence of such load errors in emissions and
fuel economy measurements are unknown.
It is noted that a study by Honda Motor Company,
Ltd., which was submitted to the EPA on March 29, 1974,
found good correlation between road-loads derived from
coast-down tests and chassis dynamometer loads performed
with a fixed exponent type of hydrokinetic - closed flow
PAU. These results tend to indicate that a more comprehensive
study of tire-to-roll losses might show that sufficiently
accurate duplication of road-loads for the EPA tests could
be obtained without the use of algorithm type of controller.
It is not feasible to use hydrokinetic - closed
flow type PAU's for inertia simulation because they lack
motoring capability.
Clayton Manufacturing Company furnished information
on their Part No. D-18773 Power Absorption Unit. Its rotor
-2 2
inertia is 2.221 x 10 kg.m , and the cooling water flow
-4 3
required is 2.09 x 10 m /skw. Price of the basic unit is
$2,145, less controls.
The unique capability of the hydrokinetic - closed
flow type of PAU to "lock up" a load index point on its
built-in exponential characteristic curve is advantageous
when it is desired to operate on a fixed curve of speed
versus load. However, it would be necessary to add and
withdraw fluid from the amount initially locked up in the
circulating loop to index the basic load in order to correct
for the constant and linear terms in the load algorithm
furnished by the EPA. This technique has been used in
hydrokinetic PAU's for testing aircraft turboprop engines.
2-33
-------
TABLE 2-4
HARLEY-DAVIDSON - 1200 CC
KPH
20
40
50
65
85
110
20
40
50
65
85
110
HORSEPOWER
NET PAU
HP -EPA
.26
.97
1.69
3.43
7.26
15.27
- .04
.58
1.26
2.84
6.31
13.43
HORSEPOWER
PAU LOAD*
2.83 EXP
.12
.87
1.63
3.43
7.33
15.20
HONDA - 90 CC**
.10
.72
1.35
2.84
6.07
12.59
HORSEPOWER
ERROR
- .14
- .10
- .06
0
+ .07
- .07
+ .14
+ .14
+ .09
0
- .24
- .84
ERROR
54
10
4
0
1
0.4
350
24
7
0
4
6
* PAU load indexed @ 65 kph.
** Used EPA data for motorcycle of similar mass to predict Honda 90 road load
2-34
-------
It is possible to re-establish the initial index load at any
time because the fluid withdrawn is held in a hydraulic
cylinder from which it can be returned to the circulating
loop. If, as seems now to be the case, the exponent of the
net PAU load versus speed (motorcycle load minus tire/roller
loss) is confirmed to vary from 2.4 to 3.1 as shown in
Figures 2-2 and 2-3, the above described means of changing
the PAU fill could be used to change the fixed exponential
characteristic of a hydrokinetic - closed flow PAU as required
to improve the accuracy of approximating the road-load
curves of individual motorcycles. This technique has been
recently demonstrated by Clayton Manufacturing. However,
this technique requires substantially more development
before it is fully automated. This possibility is suggested
because the controller would be less complicated than that
needed for algorithm-type loading, and may be substantially
less expensive than comparable eddy current and DC motor
generators.
2.7.6 Eddy Current
The eddy current PAU consists of a toothed rotor,
made of magnetically permeable material such as steel, which
rotates in close proximity of magnetized field poles. The
poles are arranged so that the lines of flux pass through
the teeth of the rotor producing a higher flux density in
the electromagnet pole face areas which are in proximity
with the rotor teeth and lower flux density in areas between
the teeth. As the rotor turns, any given point on the face
of each electromagnet pole experiences changes in flux
density, resulting in the generation of eddy currents and
local magnetic fields which react with those in the rotor
teeth. The interreaction of these magnetic fields produces
resistance to relative motion between the rotor and stator
or torque.
2-35
-------
The magnitude of this torque is directly related
to the flux density which, in turn, is a function of the
current in the coils of the field. The eddy currents in the
pole faces generate heat which is dissipated either by
passing cooling water within the pole faces or by injecting
water directly into the gap between the rotor and stator.
The former technique is usually employed in eddy current
PAU's of smaller capacity in which the additional hydrodynamic
drag is objectionable. Water-in-gap construction is usually
preferred in large capacity PAU's because the added drag is
acceptable and it permits a more compact size.
The eddy current PAU marketed by Lear Siegler
differs from the above in that it is air-cooled and eddy
currents are generated in disc rotors at each end of an
axial electromagnet assembly. These rotors are finned or
slotted to facilitate dissipation of heat.
The basic speed/load characteristics of the eddy
current PAU are not acceptable for the motorcycle dynamometer
facility because the torque produced at a constant field
excitation current tends to be relatively constant. However,
highly developed, solid-state closed loop controls capable
of following an externally generated electrical signal
representing the polynomial speed/load algorithm are available.
In addition to the absorption-only eddy current
PAU's, there are universal models available in which an AC
induction motor and eddy current clutch are packaged together
with the eddy current absorber. These universal models
provide motoring capability and could be used for vehicle
inertia simulation.
The Dynamatic Division of Eaton Corporation has
provided information on their Model 758 DG PAU, rated at
36.8 kw (50 HPm), which is for absorption only. Catalog
price of the basic mechanical unit is $6,560. The standard
package which also includes controller, torque measuring
system, digital tachometer, calibration arms, etc., is
2-36
-------
priced at $12,260. Their universal Model A15U, which has
about the same absorption capacity, plus 11.0 kw (15 HPm)
motoring is priced at $23,935 for the basic machine. This
would be required for PAU inertia simulation.
Meidensha Electric Manufacturing Company, Ltd. has
proposed their Model TW-55 for motorcycle testing. The
moment of inertia of the rotor is 4.7 x 10"1 kg.m2. Cooling
water required is given as 1.00 x 10"4 m/3s. The price of
this PAU, including solid-state power supply and controller
is $15,000.
Lear Siegler, Inc., Industrial Electrical Products
Division, also provided literature. Their power absorber is
of the eddy current type, but differs in that discs, in
which eddy current heating takes place, are air cooled. The
windage loss, although probably small at the operational
speeds contemplated, is not measured. Their Model C-40
seems to have the necessary capacity for the motorcycle
dynamometer facility. The rotor inertia is 3.531 x
1 2
10 kg.m . Price of the basic absorber is $2,100. Although
Lear Siegler offers a control for this absorber, it is
unsuitable for the motorcycle PAU application because it
provides only manual excitation control or automatic constant
speed control. A more complex control having an output of
10 amperes at 90 volts, DC would be required and is available
from Lebow Associates.
Lear Siegler does not offer a PAU with combined
motoring and absorbing capability.
2.7.7 DC Motor/Generator
A DC-type PAU connected to the roller would provide
means of not only absorbing power, but also driving the
roller as required for simulating vehicle inertia without
the use of flywheels. Generally, DC-type PAU systems, when
operating in power absorbing modes, are capable of faster
2-37
-------
response to program inputs and more precise speed regulation
than eddy current PAU's. However, they typically are much
higher in cost, have an inherently greater moment of inertia,
require more complex controls and are more expensive to
maintain than universal-type eddy current PAU's.
Meidensha Electric Manufacturing Company, Ltd.
submitted information on three DC-type PAU's. All are
intended to be used for motorcycle dynamometer systems in
which part of the vehicle weight simulation is performed by
means of the PAU and controller. The DC-type PAU's range in
price from $24,200 to $26,850, including the power supply,
controller and some ancillary equipment. Meidensha submitted
detailed price information for chassis emission dynamometers
and durability units. These are included as Tables 2-5 and
2-6.
Reliance Electric Company suggested their 14.9 kw
(20.3 HPm) DC PAU, Type No. 2810, and Controller Type No. S-3R
The approximate price of these components is $20,000.
Delivery is currently 35 to 40 weeks.
The Reliance PAU is not trunnion mounted. Torque
measurement would require the addition of a brushless-type
torque sensor or a torque table which would add $1,500 to
$3,000 in cost.
2.7.8 AC Adjustable-Speed Drives
AC induction motors can be operated over a wide
speed range by connecting them to a source of variable
frequency power. The resulting drive has the advantages
when compared to DC drives of lower maintenance due to the
elimination of brushes and commutator and speed control'
without tachometer feedback.
Recent developments in solid-state devices have
made such drives more practical in the following way. The
basic system for variable frequency AC power supply consists
2-38
-------
Table 2-5. PRICE AND SPECIFICATION FOR EMISSION DYNAMOMETER
DC Chassis Dynamometer
with Flywheel only
Total Price: $48,150
DC Chassis Dynamometer
with two flywheels and
70 kg Electrical
Inertia Simulation
Total Price: $38,950
DC Chassis Dynamometer
with one Flywheel and
150 kg Electrical
Inertia Simulation
Total Price: $33,950
DC Chassis Dynamometer
with 300 kg Electrical
Inertia Simulation
Total Price: $31,270
Eddy Current Chasiss
Dynamometer with
Flywheel only
Total Price: $38,950
ROLLER
Steel made, 530.5mm
in diameter - (ONE)
$3,650
n
$3,650
n
$3,650
M
$3,650
II
$3,650
FLYWHEEL
10 kg + 20 kg + 40 kg
+ 80 kg + 150 kg
direct control
$20,300
80 kg + 150 kg
direct control
$10,000
150 kg remote
control
$4,600
Fixed Flywheel 180 kg
including roller and
dynamometer inertia
mounted on roller
shaft
$770
10 kg + 20 kg + 40 kg
+ 80 kg + 50 kg
direct control
$20,300
DYNAMOMETER WITH
CONTROLLER
DC Cradled Dynaometer
with Controller
15 hp absorption,
11 hp motoring
(1,000 r om- 100 km/h)
$24,200
n
$25,300
DC Cradled Dynamometer
with Controller
17 hp asborption,
13 hp motoring
1,000 rpm (100 km/h)
$25,700
DC Cradled Dynamometer
with Control
20 hp absorption,
15 hp motoring
100 rom (100 km/h)
$26,850
15 hp absorption
1,000 rpm model TW-55
cooling water 6L/min.
$15,000
Prices of Roller and Flywheel are for your information.
Above price is based on F.O.B. at Yokohama and included
boxing charge.
2-39
-------
Table 2-6. PRICE AND SPECIFICATION FOR DURABILITY DYNAMOMETER
DC Chassis Dynamometer
with Flywheel only
Total Price: $50,050
DC Chassis Dynamometer
with two flywheels and
70 kg Electrical
Inertia Simulation
Total Price: $40,500'
DC Chassis Dynamometer
with one Flywheel and
150 kg Electrical
Inertia Simulation
Total Price: $40,150
DC Chassis Dynamometer
with 300 kg Electrical
Inertia Simulation
Total Price: $42,820
Eddy Current
Dynamometer with
Flywheel only
Total Price: $38,950
ROLLER
Steel made, 530.5mm
in diameter
$3,650
It
$3,650
ii
$3,650
n
$3,650
II
$3,650
FLYWHEEL
10 kg + 20 kg + 40 kg
+ 80 kg + 150 kg
direct control
$20,300
80 kg + 150 kg
direct control
$10,000
150 kg remote
control
$4,600
180 kg including
roller and dynameter
inertia
$770
10 kg + 20 kg + 40 kg
+ 80 kg + 50 kg
direct control
$20,300
DYNAMOMETER WITH
CONTROLLER
DC Cradled Dynaometer
with Controller
25 hp absorption,
19 hp motoring
(1,200 rpm-120 km/h)
$26,100
II
$26,850
DC Cradled Dynamometer
with Controller
45 hp asborption,
35 hp motoring
1,200 rpm (120 km/h)
$31,900
DC Cradled Dynamometer
with Controller
50 hp absorption,
37 hp motoring
750/1,200 rpm
(75/120 km/h)
$38,400
Eddy Current Cradled
Dynamometer with
Controller
25 hp absorption,
1,200 rpm MODEL TW-55
cooling water 9L/min.
$15,000
Prices of Roller and Flywheel are for your information.
Above price is based on F.O.B. at Yokohama and included
boxing charge.
2-40
-------
of converting the 60-cycle power to DC and then converting
the DC back to AC by switching or "chopping" at the desired
frequency. Early variable-frequency systems were limited by
the difficulty of turning off the conduction in silicon-
controlled retifiers (SCR's) which were then the most prac-
tical high power solid-state switching devices. The recent
development of a duplex transistor in a single chip, called
the Darlington, has eliminated the complex circuitry which
was required to turn off the SCR's. Darlingtons are capable
of higher frequency operation so that adverse heating effect
of square wave operation of motors can be reduced by six-
step square wave pulse-width modulation.
It appears that adjustable-speed AC drives are
advantageous in applications wherein sparkless operation,
ultra-high speed, multiple-drive synchronization and/or
adverse environmental conditions are a factor. None of
these seem to be considerations in the selection of a PALI
for the motorcycle test facility.
2.8 POWER ABSORPTION SYSTEMS - RANKINGS
2.8.1 Ranking Alternatives - Flywheel Inertia Simulation
Optimum Selection
1. Eddy Current-type PAU, absorption only,
control for polynomial speed/load algorithm.
2. Hydrokinetic - closed flow-type PAU, absorption
only, fluid displacer and control to correct
basic exponential speed/load characteristic
to polynomial speed/load algorithm.
2-41
-------
Acceptable Selection
3. Hydro kinetic - open flow-type PAU, absorption
only, control for polynomial speed/load
algorithm.
4. DC motor/generator.
Unacceptable Selection
5. Air blower.
6. Hydraulic pump.
2.8.-2 Ranking Alternatives - Partial PAU Simulation
Of Inertia
Optimum Selection
1. DC motor/generator-type PAU, motor and absorb,
control for polynomial speed/load algorithm
plus inertia simulation.
2. Hydraulic pump/motor-type PAU, control for
polynomial speed/load algorithm plus inertia
simulation.
Acceptable Selection
3. Universal eddy current drive, motor and
absorb, control for polynomial speed/load
algorithm plus inertia simulation.
2-42
-------
Not Appliable
4. Hydro kinetic - closed flow-type PAU.
5. Hydrokinetic - open flow-type PAU.
6. Air blower.
2.8.3 Discussion
Final selection of two, or possibly three, of the
alternative systems should be made based upon the following
considerations:
• Accuracy required in duplication of motorcycle
road-loads to assure equitable emissions and
fuel consumption measurements.
t Evaluation of proven, available control
systems, versus the probably time and cost
for development of more elaborate systems.
• Probability of changing test procedures and
consequent adaptability required to avoid
obsolescence of equipment.
These selections should then be restudied, obtaining detailed
specifications, prices and deliveries from vendors from
which to make final choices.
2.8.3.1 PAU's with Flywheel Inertia Simulation
Optimum
The study to date, of available information, tends
to indicate that an eddy current- type PAU in combination
with a mechanical flywheel weight simulation system and a
2-43
-------
stock, high performance-type industrial drive controller,
plus a computer to generate the control command from the
speed/load algorithm is the optimum system for accurate
duplication of a wide range of motorcycle performance charac-
teristics, with a minimum of development time and cost.
The hydrokinetic - closed flow-type PAU, with a
fixed exponential speed/load curve, in combination with
tire-to-roller interaction would be the simplest and lowest
cost system. Differences in actual road operational loads,
versus the dynamometer loads, need to be evaluated in terms
of their influence on emissions and fuel economy measurements
to determine if they are significant compared to other
uncertainties such as the road data, etc. It is possible to
improve the accuracy of the hydrokinetic - closed flow:type
PAU's speed/load fit by means of a fluid displacer to correct
the basic exponential curve of a hydrokinetic - closed flow
unit to the polynomial algorithm by changing the volume of
fluid in the working loop. However, a detailed analysis,
based upon precise definitions of accuracy and response time
requirements, would be necessary to determine the feasibility
of thi s solution.
Acceptable
The hydrokinetic - open flow-type PAU would be
acceptable for either fixed exponential or polynomial speed/
load testing. However, because the open flow system requires
matching of input to output flow, a moderately expensive
control system would be required, even for the simpler
exponential .loading. The control would have to monitor
speed and torque and continuously adjust the inlet and
outlet flow to maintain the required fill. This would
result in a cost disadvantage compared to a hydrokinetic -
closed flow PAU system for equal performance.
2-44
-------
The addition of a polynomial command signal would
provide the ability to more accurately duplicate the road-
load, although it is doubtful if the rate of response could
be made equal to that of the eddy current system. This
system would require engineering development costs greater
than those for the eddy current system.
The DC motor/generator, while acceptable, is not
optimum because of higher cost. The motoring capability and
the faster response to control command signals may be useful
with a flywheel inertia simulation system, and might warrant
use of this higher cost system.
Unacceptable
The air blower is unacceptable because of its
size, its inertia and the complex damper system needed to
match loads and power curves.
The hydraulic pump PAU could be used for the poly-
nomial loading requirement. However, the several disadvantages
discussed in the test render it unacceptable. The system
would require a reservoir, close temperature control heat
exchanger to minimize viscosity changes, and a flow control
system to produce an approximately cubic speed/load character-
istic. Maintenance of filters, etc., would be greater than
for a PAU using water as the working fluid.
2.8.3.2 PAU's with Inertia Simulation
Optimurn
The DC motor/generator is believed to be the best
choice for a dynamometer system in which vehicle inertia is
simulated by one to four flywheels plus programmed motoring
and braking by the PAU to "trim" the inertia to the 10 kg
increments of vehicle mass specified in the NPRM. (This
2-45
-------
technique is discussed in the Task 4 report.) Primary
reasons for ranking the DC motor/generator first are that
these machines and controls have been used in complex indus-
trial drives successfully. They would require a minimum of
development for adaptation to the motorcycle test facility
requi rement.
The hydraulic pump/motor and control system has
also been used in industrial drives, although, perhaps, not
as extensively as the DC motor generator. It is possible
that a more detailed dynamometer facility performance speci-
fication would result in ranking the hydraulic pump/motor
above the DC motor/generator on the basis of fast response
time. However, development and maintenance costs would be
higher.
Acceptable
Universal eddy current drives are ranked below DC
motor/generators and hydraulic pump/motor systems primarily
because of their slower response. The acceptability of a
slower response to the load command signal depends, ultimately,
on how much emissions and fuel economy measurements are
affected.
Not Applicable
The hydrokinetic open and closed flow and air
blower types of PAU's are not applicable because they lack
motoring ability.
2-46
-------
Section 3
TASK 3 - EVALUATION OF ROLL CONFIGURATIONS
3.1 INTRODUCTION
This task report will present an analysis of the
drive roll assembly for use in the motorcycle chassis dyna-
mometer facility. Performance requirements defined in the
EPA Notice of Proposed Rulemaking (NPRM) and Contract
No. 68-03-3241 will be examined, and a general specification
for the drive roll assembly will be presented. A detailed
analysis of the effects of various rol.ler design parameters
is presented, and recommendations are made in Section 3.8.
In Section 3.3 experiments to quantify the tire/roll inter-
actions will be described and quantitative data will be
presented and used to support the design decisions.
3.2 REVIEW OF REQUIREMENTS
The NPRM defines the driving cycle for emissions
measurement and durability testing. Both procedures require
vehicle operation at idle, steady-state, acceleration and
deceleration modes. The maximum speed in the emissions
procedure is 90.9 kph; 110 kph (lap 11) in the durability
cycle. The power to be transmitted from the motorcycle tire
through the roller to the power absorption unit (PAU) has
been examined in Section 2. Equations and empirical load
factors for several motorcycles were developed in Table 2-1.
3-1
-------
This data has been used to select speed/power points for
tire to roller tests as discussed in Section 3.3 of this
report.
The contract indicates that cradle rolls of 0.203
to 0.229 m (8 to 9 in.) diameter and spaced 0.457 m (18 in.)
apart, which have been used in automobile chassis dyna-
mometers, cause undesirable tire flexing, and the use of a
single roller of larger diameter is recommended by the EPA
for motorcycle testing. This report will quantitatively
examine these alternatives. Consideration of the effects of
surface texture and crowning of the roller is also included
in the design study.
The contract emphasizes the importance of minimum
friction and drag in the roller assembly. This is parti-
cularly important when testing smaller motorcycles which
produce low power outputs. Also, the characteristics of the
dynamometer should not be affected by temperature fluctuations
and time.
3.3 QUANTIFICATION OF TIRE/ROLL LOSSES
Although the EPA contract indicates that a large
diameter, single roller is desirable to minimize tire flex-
ing, little data is available quantifying tire-to-roller
interaction for motorcycles. Since these losses plus the
PAU and inertia loads constitute the total load on the
motorcycle being tested, it was essential that they be
measured. An experiment was, therefore, designed to obtain
these data as a function of roller size, roll arrangement
(single roller or cradle rollers) and motorcycle weight.
3-2
-------
3.3.1 Test Description
Three roller diameters, 0.217 m (8.65 in.), 0.324 m
(12.75 in.), and 0.508 tn (20.0 in.) were fabricated and
tested. The two smallest diameters were tested in both
cradle and single roller configurations. Only single roller
tests were made with the largest roller. The roll centerline
spacing was 0.395 m (15.56 in.) for the 0.217 m (8.65 in.)
diameter rolls and 0.449 m (17.69 in.) for the 0.324 m
(12.75 in.) rolls, which resulted in approximately equal
loadings on each of the rolls.
Two motorcycles were used. The larger one, a
Harley Davidson FLH, was equipped with a 5.10-16 size tire
inflated to 179 KPa (26 psig) pressure cold. The rear axle
load was 268.5 kg (592 Ibs.). The smaller motorcycle, a
Honda 90cc unit, was equipped with a 2.75-17 size tire
inflated to 96.5 KPa (14 psig) and 138 KPa (20 psig) cold.
The rear axle load was 107.5 kg (237 Ibs.). Initial tests
were made on the smaller motorcyle with the tire inflated to
96.5 KPa (14 psig), and excessive tire deflection was observed
operating on a single 0.217 m (8.65 in.) diameter roller. A
second complete test series was run using 138 KPa ,2Q psig)
tire pressure. No tests were performed on the larger motor-
cycle using reduced tire pressure.
The rollers were carried on antifriction bearings
without seals or shields and lubricated with automatic
transmission fluid to minimize friction. Adjustable loading
was provided by a manually-indexed hydrokinetic-type PAU to
ensure stability and for convenience in operation. This
type of PAU has an inherent exponential speed/load character-
istic; whereby, absorbed power changes approximately as the
2.8 power of rpm. The initial set point or index is deter-
mined by the amount of fluid "locked up" in the unit. The
speed/load curve remains constant until the amount of fluid
in the recirculating loop is changed. Two loading indices
3-3
-------
for each of the motorcycles were chosen. This selection was
based upon an approximation of the road-load predicted by
the EPA equation and a higher load representing added power
for acceleration. The load index is identified in the
plots. The higher letter is used to denote a higher load
index; i.e., load "E" is greater than load "D".
Motorcycle rear wheel and PAU torques were measured
by means of strain gauge load cells. Motorcycle wheel rpm
was measured by an electromagnetic pulse generator and an
electronic counter. PAU rpm was measured by means of a DC
tachometer generator and a digital voltmeter.
Operating test points were selected to be representa^
tive of the speeds and loads appropriate for the EPA emissions
test procedure for motorcycles of the two mass classes.
Approximately the same speed/load points were run on the
following roller setups.
1. Two 0.217 m (8.65 in.) diameter rolls on
0.395 m (15.56 in.) centers.
2. One 0.217 m (8.65 in.) diameter roll.
3. Two 0.324 m (12.75) in.) diameter rolls on
0.449 m (17.69 in.) centers.
4. One 0.324 m (12.75 in.) diameter roll.
5. One 0.508 m (20.0 in.) diameter roll.
3.3.2 Roller Test Results
Laboratory test data are presented in Appendix A
and several figures which shall be discussed individually.
Torque losses were determined by measuring independently the
input torque and PAU, and determining the difference.
3-4
-------
Curves through the data points were established using
logarithmic curve fit techniques.
It was confirmed that losses in transmitting
torque from the rear wheel to the roller become smaller as
the roller diameter is increased. This is illustrated in
Figure 3-1, a plot of single roller tests on the Honda 90
motorcycle. In this plot, data is provided for two roller
diameters and two load indices; the higher indicated by the
suffix letter "G".
It can be observed that the torque loss change for
a given change in input torque is much smaller with a 508 mm
(20 in.) roller than it is with a 200 mm (8.65 in.) roller.
For example, increasing the input torque from 20.3 nt m to
47.5 nt m (15.0 ft. Ibs. to 35.0 ft. Ibs.) on the 20.00 G
load curve results in the torque loss increasing from 7.5 nt m
(5.5 ft. Ibs.) to 8.2 nt m (6.0 ft. Ibs.), or a change of
0.7 nt m (0.5 ft. Ibs). The same change on the 8.65 G
roller results in the torque loss increasing from 15.2 nt m
to 34.2 nt m (11.2 to 25.2 ft. Ibs.), a change about 32 times
as large.
The difference in slope of the "G" and "B" curves
indicates that the power level influences the torque loss,
possibly due to heating of the tire tread and carcass, but
investigation of this was not feasible because of time and
budget limitations.
Figure 3-2 shows the torque loss relative to
motorcycle wheel speed. The torque losses are observed to
increase a greater amount for the smaller roller per unit
increase in speed than for the larger roller as was expected.
The small difference in slope between the 20.00 B
and 20.00 G curves are noteworthy because torque losses as a
function of increased power at a given speed are nearly
constant. This is in contrast to the data shown in Figure 3-2
for the smaller roll in curves 8.65 B and G. There, data
tend to indicate that the larger, 508 mm (20.0 in.), diameter
3-5
-------
TIRE/ROLL LOSSES VS INPUT TORQUE
40
!
Z
ill
i
0
D
'
)
•l
i
Z
I!
'
)
Q
2
36
32
28
24
20
16
12
220B
220G
10
20 30 40 50
MC INPUT TORQUE (NT-M)
60
Figure 3-1. Honda Motorcycle Data Single Roll
3-6
-------
TIRE/ROLL LOSSES VS WHEEL SPEED
-
•
Z
'
,,
h
i
.
2
36
32
28
24
20
220G
16
12
220B
508B
508G
1
10
20
30 40
MC WHEEL SPEED (RPMxIO)
50
60
Fifura 3-2. Honda Motorcyclt Data Singto Roll
3-7
-------
roller would result in more nearly constant tire/roller
losses with changes in load and speed than one of smaller
diameter.
Results of operating the Harley Davidson FLH
motorcycle on single rolls of 220 mm (8.65 in.) and 324 mm
(12.75 in.) diameter are plotted on Figures 3-3 and 3-4
respectively. Again, it is observed that the torque losses
are reduced by increasing the roller diameter.
These curves indicate that the change in torque
loss per unit change in motorcycle input torque is reduced
by increasing the roller diameter. Also, increased motorcycle
input torque results in smaller changes in torque loss per
unit change in input torque as indicated by the "E" curves
and the "D" curves in Figures 3-3 and 3-4. This is thought
to be the results of increased tire temperature, although no
tire temperature data was recorded to substantiate this.
The relationship between roller diameter, speed
and torque loss for the larger motorcycle is plotted on
Figure 3-4. The larger roller was again observed to minimize
the change in torque loss for a unit change in speed. These
data tend to indicate that the 508 mm (20.00 in.) roller is
superior to the 220 mm (8.65 in.) or 324 mm (12.75 in.)
rollers because it results in minimum torque loss at typical
input torques and operating speeds. It is also believed
that the 508 mm (20.00 in.) diameter is probably near optimum,
if intrinsic inertia, installation space and cost are con-
sidered. However, much more extensive experimentation would
be necessary to adequately support this opinion. Such
experiments should include a large variety of motorcycles
and axle loads, additional tire types and sizes and more
test points per combination.
The comparative performance of single roll and
cradle roll arrangements can be seen in Figures 3-5 and
3-6. As in the data obtained with the single roll, losses
decrease as roll size increases. Although only limited data
3-8
-------
140.
TIRE/ROLL LOSSES VS TORQUE INPUT
120
100
220E
i
.
u
I
Q
)
'
•1
.
1/3
1
2
01
o
a
'
.
80
60
40
20
324D
220D
•;
508E
40
80
"12ZT
"205"
. -1
MC INPUT TORQUE (NT • M)
Fifurt3-3. H«r1«y DivMton Motorcydi Single Roll
3-9
-------
140.,
TIRE/ROLL LOSSES VS WHEEL SPEED
120
100
:
-
80
O
i
•
-
2
60
a
a
40
20
1
L
10 20 30 40 50
MC WHEEL SPEED (RPMX 10)
1)0
70
80
90
Figure 3-4. Harley Davidson Motorcycle Single Roll
3-10
-------
CRADLE VS SINGLE ROLL
40
te
2 26
:
'
>
I
Z 20
5
u
i
-
16
>
• i
•
10
324B&C
CRADLE ROLL
SINGLE ROLL
20
30 40
50-
INPUT TORQUE(NT" M)
60
. I
70
SO
3-6. Honda Motorcyclt Data
3-11
-------
140
CRADLE ROLLS VS SINGLE ROLL
120_
- 100
O
cc
O
-< 80
'
tr
C 60
40
20_
220D
220E,
324E
-324D
CRADLE ROLL
SINGLE ROLL
0
1
/x**
1111
i 40 80 120 160
I
200
INPUT TORQUE (NT-M)
Figur* 3-6. Harlty Davidson Motorcycle Data
3-12
-------
was obtained, it appears that the losses experienced with
single rollers are slightly less than those resulting from
use of cradle rollers.
As shown in Appendix A, a series of tests were
conducted with the Honda bike at two different tire inflation
pressures. Data was limited and no discernible relationship
between tire/roll losses and tire inflation pressures could
be established.
During the course of the test program, utilizing a
single roller configuration, it was observed that tire/roll
losses increased if the rear wheel of the motorcycle was not
properly positioned atop the roller. If the motorcycle's
rear axle was positioned either in front or behind the top
dead center of the roll the losses appeared to increase.
3.4 ROLLER SURFACE CONTOUR
Making the diameter of the roller larger in the
center than at the ends is a means of causing the tire
contact patch to be centered automatically on the roller,
despite limited skewness between the wheel and roller axes.
This construction is used in the Hartzell Mark II motorcycle
dynamometer.
The operating principle is sketched in Figure 3-7.
A wheel rolling with its axis inclined to the surface is
shown in Figure 3-7a. It will describe a circular track
with radius 'r' as shown in Figure 3-7b. The radius of its
path will be the radius of the wheel axis with respect to
the surface. For example, a 610 mm (24 in.) diameter wheel
inclined 3 degrees will roll on a circle of approximately
5800 mm (229 in.) radius.
Similarly, a wheel running on a crowned or tapered
surface roller will travel in a spiral path toward the
larger diameter of the roller as shown in Figure 3-7c.
3-13
-------
B
Figure 3-7. Roll Surface Contour
3-14
-------
Equilibrium of the forces acting on the tire contact patch
will occur at a point near the maximum diameter if, in this
position, the wheel and roller axes are parallel. The
distribution of contact area will be unequal and increased
losses due to tread creep will result if, when the tire
patch is approximately centered over the maximum roller
diameter, there is skewness between the axes. This loss
does not occur on a cylindrical roller. If the skewness
between the axes is in excess of that which can be accommodated
by the amount of taper or crown, the tire will run off of
the roller.
Automatic alignnment of the wheel and roller axes,
rather than centering of the tire contact patch, occurs on
straight cylindrical rollers as shown in Figure 3-7d. The
center of the contact patch describes a helix of decreasing
pitch until the axes are aligned. With a headfork angle of
typically 25 degree to 30 degree, motorcycles tend to reduce
the pitch of the the helix, but no instability has been
observed in operation. When the front wheel clamp is not
centered or there is frame or wheel axle misalignment, the
tire will not run in the center of a straight cylindrical
roller, but there is no adverse effect on performance.
Similar misalignment could result in the tire operating at a
considerable slip angle, although laterally centered, on a
crowned roller.
3.5 SURFACE FINISH
Roll surface finish and contamination affect the
coefficient of friction and the maximum torque which can be
transmitted between the motorcycle tire and roller. Experience
from automobile chassis dynamometer operation is probably
applicable, even though there are some differences in tire
cross section, unit loading, etc.
3-15
-------
Application of hard grit coating is of value in
preventing skidding or tire breakaway where water, oil or
dirt is present. However, there is no appreciable advantage
when the tire and roll surface are reasonably clean and dry.
The disadvantages of abrasive coating the roller are: increased
tire tread wear, which may be hazardous during prolonged
tests, and the coating of the roller with tread compound
under some circumstances. Another consideration in the use
of grit coating is that the surface gradually changes due to
dulling of the particle edges and loss of particles due to
bonding failure.
Both circumferential and axial grooving have been
tried. These grooves also tend to increase tire wear when
new and lose effectiveness as the edges of the grooves
become rounded. A combination of right and left hand spiral
grooves, resulting in an elongated diamond pattern, has
recently been developed by Clayton Manufacturing. It seems
to be quite effective in maintaining traction even with wet
or dusty tires, yet does not show evidence of tread wear or
tendency to become coated with rubber.
3.6 MATERIAL OF CONSTRUCTION - ROLLER
Aluminum, with an anodic surface treatment, has
been used for roller fabrication. Although it is reputed to
have a higher coefficient of friction in combination with
tire tread stock than that for steel, it is doubtful if
there is a demonstrable, practical improvement in actual
operation. There is little to be gained in reduced roller
inertia since an emissions test dynamometer must simulate
the inertia of the vehicle.
3-16
-------
3.7 ROLLER MOUNTING
The roller shaft bearings should be selected to
provide a B-10 life of 10,000 hours or more, as defined by
the Anti-Friction Bearing Manufacturers Association, to
assure reliability and minimize maintenance. The use of low
friction seals or shields and oil lubrication would result
in low and stable drag losses, which are desirable for the
contemplated services.
3.8 RECOMMENDATIONS
Based upon the results of laboratory data and
detailed discussion of design features in Section 3 of this
report, the following roller specifications are proposed.
Type - single roller
Diameter - 530.5 mm (20.9 in.)
Width - 300 mm (12 in.)
Material - low carbon steel
Surface - cylindrical (uncrowned)
Surface texture- smooth
The limited tests performed are believed adequate
to be insure that the roller specified above will operate
satisfactorily in the contemplated motorcycle dynamometer.
3-17
-------
Section 4
TASK 4 - EVALUATION OF INERTIA SIMULATION METHODS
4.1 INTRODUCTION
This section will present an analysis of methods
for the simulation of vehicle inertia with a chasiss dynamom-
eter. The EPA Proposed Rulemaking (NPRM) and Contract No.
68-03-2141 define inertia requirements, and these are examined
A general specification for the inertia simulation system is
presented. An analysis of the two basic methods available
for inertial simulation is detailed in Section 4.4 of this
report and recommendations are made in Section 4.5.
4.2 REVIEW OF REQUIREMENTS
The NPRM defines the driving cycle for emissions
measurements in Appendix I and the durability schedule in
Appendix IV. Inertia requirements for both cycles are the
same, but depending on the simulation technique, they may
impose different specifications for meeting the requirements
of the two cycles. The contract stipulates that if flywheels
are used for inertia simulation, then they shall be coupled
to the roller automatically by means of remote control such
as electric switches or pushbuttons, and that such control
shall be fail-safe to prevent injury to the operator or
damage to equipment in the event of loss of electrical
power, air supply, etc. A further requirement is that if
4-1
-------
flywheels are driven through a speed increaser, such as a
gearbox, then it shall be a low-friction device to avoid
inaccuracy of vehicle loading resulting from unmeasured
losses in the drive.
Increments of 10 kg from 100 to 700 kg loaded-
vehicle mass are specified in the contract with consideration
to be given to reducing this maximum from 700 to 500 kg as a
means of reducing system complexity and/or improving accuracy
or precision. Similarly, the size of the minimum increment
influences the system performance and cost. The possibility
of using smaller increments such as 5 kg, for smaller vehicles
especially, to improve accuracy and the use of larger,
possibly 20 kg, increments to reduce complexity, should be
evaluated.
It is indicated that both the use of multiple
flywheels and/or the use of programmed motoring and absorbing
by the power absorption unit (PALI), are to be considered for
vehicle inertia simulation.
Neither the contract or the NPRM indicate the
permissible tolerance on accuracy or response time for the
inertia simulation system. However, emissions will be
influenced by the vehicle inertia simulation accuracy, since
the horsepower which must be developed at the driving wheel
during acceleration is directly affected by the inertia.
4.3 GENERAL SPECIFICATIONS
In light of the requirements for emissions tests
as proposed in the NPRM and the contract, a basic set of
minimum specifications can be established. These are listed
in Table 4-1 and 4-2. Two specifications are called out
because of the differences imposed by utilizing mechanical
or electrical simulation.
4-2
-------
Table 4-1. GENERAL SPECIFICATION FLYWHEEL INERTIA SYSTEM
Range
Accuracy
Control
Function
Interlock
Response Time
Windage and Friction
Losses
100 to 700 kg equivalent vehicle weight in 10 kg incre-
ments with optional reduction to 500 kg.
Plus or minus 1.5 percent of selected value.
Remote switches or pushbuttons.
Fail safe in case of loss of electric power, air
pressure, etc.
Range change blocked out above zero rpm in flywheel
system.
Not applicable to flywheel system.
Inertia simulation system not to introduce error
greater than 5 percent of road load force at any
operating point.
Table 4-2. GENERAL SPECIFICATION ELECTRICAL INERTIAL SIMULATION
Range
Additional PAU
absorption capa-
bility
PAU motoring
capability
Accuracy
Control
Function
Interlock
Response Time
Windage and Friction
Losses
100 to 700 kg equivalent vehicle weight in 10 kg incre-
ment with optional reduction to 500 kg.
18.4 kw (25 HPm).
18.4 kw {25 HPm).
Plus or minus 1.5 percent selected value at any speed
above 5 kph and 5 percent from 1 to 5 kph.
Remote switches of pushbuttons.
Fail safe in case of loss of electric power, air
pressure etc.
Range change blocked out above zero rpm.
0.5 sec. for 90 percent of ,step change in PAU type
simulation system as measured in force change.
Inertia simulation system not to introduce error
greater than 5 percent of road load at any operating
point.
4-3
-------
4.4 INERTIA SIMULATION SYSTEM CHARACTERISTICS
In this section, the performance, suitability and
probable cost of mechanical-type and PAU-type inertia simula-
tion systems shall be compared as they relate to the require-
ments of motorcycle emissions testing and durability driving
cycles. Design and operation characteristics shall be
considered as follows:
• Calibration
• Accuracy of simulation
t Response to change in rpm
t Stability of calibration
• Convenience
• Maintenance
The primary consideration in comparing inertia
simulation systems shall be the emissions test procedure as
defined in the EPA NPRM and Contract No. 68-03-2141.
Durability driving requirements shall be regarded as secondary
4.4.1 Mechanical Inertia Simulation System
The flywheel is a primary standard of inertia. As
proposed for this application, the unit consists of coupling
combinations of flywheels to the common shaft of the roller
and PAU so that the combined inertia is equivalent to that
which is required; i.e., 100 to 700 kg in 10 kg steps. This
system would consist of a fixed flywheel which, in combination
with the PAU and drive roller inertias, would be equivalent
to 100 kg vehicle inertia weight. Additional flywheels of
10, 20, 40, 80, 160 and 320 kg equivalent vehicle inertia
would be engaged singly or in combination to provide the 60
increments required.
4-4
-------
The mechanical inertia system is constructed in
the following manner. A drive shaft, which is coupled to
the roller shaft and one fixed inertia wheel, is carried on
antifriction bearings. The additional flywheels are carried
on antifriction bearings and each can be individually coupled
to the shaft by means of a latching-type air-operated clutch.
Clutch actuation and selection of the flywheels is controlled
from the operator's station by means of electrical switches
or pushbuttons. An electrical interlock is provided so that
the clutches can neither be engaged or disengaged when the
flywheels are in motion.
Maintenance, consisting of lubrication, adjustment
of clutches and electropneumatic actuators would not be
appreciably more than necessary for the PAD and roller.
The windage, bearing friction and seal drag of the
flywheels and roller can approach the total load of a small
motorcycle. This unmeasured loss varies with the number of
flywheels engaged and with the operating rpm. However, this
loss is not expected to change as a function of system
usage. A value of 4 Nt.m (3 Ibs. ft.) torque, rising slightly
with rpm is expected. This loss has been estimated based on
the performance of existing dynamometers used in the testing
of light-duty vehicles. Data on those systems indicate that
system drag characteristics are quite stable after initial
break-in. This loss is equivalent to the torque exerted by
the rear wheel of a Honda 90cc unit rotating at 20 rpm.
Although it can be minimized by careful design, elimination
of this variable error is desirable. This can be done by
connecting a small (3 kw) motor to the shaft, programmed to
apply driving torque to compensate for the windage and
frictional losses of the flywheels engaged at the instanta-
neous speed of operation. Such a compensating motor and
control could be used in conjuction with PAU's incapable of
motoring, or the PAU absorption characteristics could be
programmed to take into account this inherent loss. PAU's
4-5
-------
having motoring capability could be utilized for compensating
flywheel losses by modifying the calculation of road-load
and acceleration force as required. It is expected that
sufficient accuracy could be obtained with a simple control.
The flywheel windage and friction compensation calibration
could be checked by driving the roller up to a given rpm and
observing whether the rpm increased or decreased during
coasting.
Since the flywheel is a primary standard of inertia,
it requires no periodic calibration checks unless it is
damaged. The response to a change in rpm is instantaneous
for all practical considerations. This is important since
any perceptible time lag could adversely affect the accuracy
of fuel consumption and emissions determinations, particularly
in those portions of the test cycle which involve gear
changing or braking.
The flywheel inertia system has greater mechanical
complexity and requires more space for installation. Mainte-
nance, consisting of lubrication, clutch service, etc.,
would not add significantly to that required on the PAU and
roller.
The use of a speed-increasing device, such as a
gearbox, between the low-speed roller shaft and the flywheel
system is attractive as a means of reducing the size of the
flywheels, bearing and clutches. There are two negative
factors to be considered. One is the loss in the transmission
of power due to pumping losses in the gears and bearings and
due to seal friction, etc. These losses can be sizeable
compared to the inertial power-loading appropriate for small
motorcycles, and they are difficult to calibrate out of the
system or otherwise compensate for because they vary with
temperature of the lubricant, gear-loading and length of
service. The second negative factor is that the added cost
of a gearbox may be greater than that of increasing the size
of the inertia simulation system components.
4-6
-------
The difference in inertial simulation system
complexity, resulting from changing the 10 kg increments to
5 kg or 20 kg increments, is relatively small in terms of
mechanical system; or the number of programming selector
switch positions for a PAU inertia simulation system would
change, resulting in the addition or deletion of one binary
step increasing the number of inertia steps to 120 or reducing
the number to 30. The change in cost is estimated to be
less than $8,000 for mechanical system and much less for a
PAU inertia simulation system. However, it is not possible
to assess the effect on emissions and fuel economy performance
of either change at this time.
4.4.2 PAU Inertia Simulation
Using a PAU having both motoring and driving
capability programmed to supplement the intrinsic inertia of
the roller, etc., is an attractive method for the emission
test cycle inertia simulation because it eliminates the need
for a complex flywheel system and provides closer simulation
than 10 kg increments. In this system, an on-line calculation
of the sum of the forces at the rear wheel due to road-load
at the instantaneous velocity, and the instantaneous rate of
change in velocity is made. The force is adjusted for the
intrinsic inertia of the roller, PAU, etc., and then this
net instantaneous force is compared to the observed force
(PAU torque) and the appropriate error signal causes the PAU
to drive or brake as necessary to make the calculated and
measured forces (torque) equal. The instantaneous force
equation is:
FT ' Fo + Flv * F2v2 + • 3T (4-1)
4-7
-------
where
FQ, Fj and F2 =• empirical factors defined by EPA data
V = velocity in meters/second
m = vehicle mass in kg
t * time in seconds
In the PAU system of inertia simulation, a digital
or analog computer calculates in real time the roller torque
or load which is appropriate for the instantaneous speed and
change in speed based upon vehicle mass. The actual torque
and the calculated torque are compared and the difference,
if any, causes the controller to increase or decrease the
PAU torque, as required. The principal difficulty with this
method of vehicle inertia simulation is that the response
time of the PAU and the roller speed-sensing system result
in some lag between the program and load. This is of less
consequence in performing the polynominal road-load curve
because the changes in PAU loading per unit time are relatively
small. However, the inertia simulation part of the control
command is determined by the rate of change of speed or
derivative which changes rapidly and is small in magnitude.
The roller speed signal can be obtained from a DC
tachometer generator or a pulse rate generator. Both exhibit
noise in the output signal; pole ripple of typically 0.5 per-
cent (peak to peak) of the signal for the DC tachometer, or
0.05 percent pitch error in the spacing of pulse generator
teeth resulting in frequency modulation of the pulse rate.
It is assumed that either generator is driven directly from
the roller shaft, but often there is additional error intro-
duced in the drive means between the tachometer generator
and roller shaft. In either case, filtering of the speed
signal must be performed to prevent the noise portion of the
tachometer output from partially.saturating the control
amplifier. This filtering introduces a time constant of,
typically, 0.25 to 0.5 seconds.
4-8
-------
In addition to the time lag due to filtering, the
PAU and controller time lag must be added. Both DC
motor/generators and universal eddy current PAU's have
considerable inductance, and the rate of change in excitation
which can be used is limited by the point at which control
instability is reached. The hydraulic pump/motor type of
PAU in combination with a sophisticated control is potentially
faster in response, but probably more expensive.
An additional concern is a system calibration. A
reliable and accurate technique for calibrating or verifying
the calibration of this system must be established. At this
time, no accurate techniques or procedures are known.
4.4.3 PAU-Type Inertia Simulation with Added Flywheel
Inertia
The addition of mechanical inertia to the system
would reduce the potential error in PAU inertia simulation
experienced by the motorcycle under test, and decrease the
maximum PAU power required. A single flywheel could be
added to make the total inertia of the dynamometer system
equal to the mean of the motorcycles to be tested, but the
time lag error in the PAU simulated portion of the total
inertia would tend to penalize vehicles of mass lower than
the mean and reduce the loads on those greater than the
mean. Use of .additional, clutchable, flywheels would further
lessen the loading errors, but would partially defeat the
benefits of using PAU inertia simulation by increasing
mechanical complexity.
4.4.4 Costs
It is difficult, at this time, to provide pricing
information about the alternative discussed. However, some
relative magnitudes and comparisons can be "guesstimated"
4-9
-------
from quotes provided by Meidensha Electric Mfg., Co., and
these have been included in Tables 4-3 and 4-4.
4.5 RECOMMENDATIONS
4.5.1 Ranking Alternatives
Optimum Selection
Multiple Flywheel system
Acceptable Selection
Single fixed flywheel with inertia corrected
command signal to motor/absorber type PAU.
Unacceptable Selection
Full electrical simulation utilizing controls
which provide command signals to a motor/
absorber type PAU.
4.5.2 Discussion
The multiple flywheel system is deemed the optimum
selection because of its inherent stability of calibration,
and the absence of response time lag. Direct-drive variable
inertia systems have been used successfully in automotive
emissions testing facilities for some time and result in
minimum losses. The use of gearboxes, cog belts or other
indirect means of coupling the flywheels to the roll, can
result in significant losses. Therefore, unless these
losses can be eliminated, indirect couplings should be
avoided.
The PAU-type of inertia system, in combination
with a single flywheel, is an acceptable selection because
4-10
-------
Table 4-3. PRICE AND SPECIFICATION FOR EMISSION DYNAMOMETER
ROLLER
DC Chassis Dynamometer Steel made, 530.5mm
with Flywhftel only in diameter - (ONE)
Total Price: $48,150 $3,650
DC Chassis Dynamometer "
with two flywheels and
70 kg Electrical
Inertia Simulation
Total Price: $38,950 $3,650
DC Chassis Dynamometer "
with one Flywheel and
150 kg Electrical
Inertia Simulation
Total Price: $33,950 $3,650
DC Chassis Dynamometer "
with 300 kg Electrical
Inertia Simulation
Total Price: $31,270 $3,650
Eddy Current Chasiss "
Dynamometer with
Flywheel only
Total Price: $38,950 $3,650
FLYWHEEL
10 kg + 20 kg + 40 kg
+ 80 kg + 150 kg
direct control
$20,300
80 kg + 150 kg
direct control
$10,000
150 kg remote
control
$4,600
Fixed Flywheel 180 kg
including roller and
dynamometer inertia
mounted on roller
shaft
$770
10 kg + 20 kg + 40 kg
+ 80 kg + 50 kg
direct control
$20,300
DYNAMOMETER WITH
CONTROLLER
DC Cradled Dynaometer
with Controller
15 hp absorption,
11 hp motoring
(1,000 rpm-100 km/h)
$24,200
n
$25,300
DC Cradled Dynamometer
with Controller
17 hp asborption,
13 hp motoring
1,000 rpm (100 km/h)
$25,700
DC Cradled Dynamometer
with Control
20 hp absorption,
15 hp motoring
100 rpm (100 km/h)
$26,850
15 hp absorption
1,000 rpm model TW-55
cooling water 6L/min.
$15,000
Prices of Roller and Flywheel are for your information.
Above price is based on F.O.B. at Yokohama and included
boxing charge.
4-11
-------
Table 4-4. PRICE AND SPECIFICATION FOR DURABILITY DYNAMOMETER
DC Chassis Dynamometer
with Flywheel only
Total Price: $50,050
DC Chassis Dynamometer
with two flywheels and
70 kg Electrical
Inertia Simulation
Total Price: $40,500
DC Chassis Dynamometer
with one Flywheel and
150 kg Electrical
Inertia Simulation
Total Price: $40,150
DC Chassis Dynamometer
with 300 kg Electrical
Inertia Simulation
Total Price: $42,820
Eddy Current
Dynamometer with
Flywheel only
Total Price: $38,950
ROLLER FLYWHEEL
Steel made, 530.5mm 10 kg + 20 kg + 40 kg
in diameter + 80 kg + 150 kg
direct control
$3,650 $20,300
80 kg + 150 kg
direct control
$3,650 $10,000
150 kg remote
control
$3,650 $4,600
180 kg including
roller and dynameter
inertia
$3,650 $770
10 kg + 20 kg + 40 kg
+ 80 kg + 50 kg
direct control
$3,650 $20,300
DYNAMOMETER WITH
CONTROLLER
DC Cradled Dynaometer
with Controller
25 hp absorption,
19 hp motoring
(1,200 rpm-120 km/h)
$26,100
II
$26,850
DC Cradled Dynamometer
with Controller
45 hp asborption,
35 hp motoring
1,200 rpm (120 km/h)
$31,900
DC Cradled Dynamometer
with Controller
50 hp absorption,
37 hp motoring
750/1,200 rpm
(75/120 km/h)
$38,400
Eddy Current Cradled
Dynamometer with
Controller
25 hp absorption,
1,200 rpm MODEL TW-55
cooling water 9L/min.
$15,000
Prices of Roller and Flywheel are for your Information.
Above price is based on F.O.B. at Yokohama and included
boxing charge.
4-12
-------
it would be a satisfactory means of inertia simulation if a
suitable PAU and control are selected, and it is determined
that the response time lag does not reduce the accuracy of
emissions, durability and fuel consumption test results to
an unacceptable degree. This approach is far superior to
the technique which utilizes full electrical inertia simula-
tion (corrected for the inherent inertia of the roll, etc.)
because of its improved response time and time lag
characteristics.
Another alternative is a system employing multiple
flywheels and motoring capability. The flywheels would
still functions as the principal method of simulating vehicle
inertia. The motoring capability could be used to perform
automated coast-down calibrations and as compensation for
the inherent losses of the system.
4-13
-------
Section 5
TASK 5 - EVALUATION OF VARIABLE FLOW COOLING SYSTEMS
5.1 INTRODUCTION
This task report will present a general analysis
of blowers and blower control systems suitable for incorpora-
tion in a variable flow blower system to provide cooling air
for motorcycle engines. The pertinent regulations which
affect the design of this system will be examined, and a
general set of specifications will be presented. A review
of blower systems currently being utilized in commercially
available motorcycle chassis dynamometers will also be
conducted. A detailed analysis of suitable blowers and flow
control methods is presented in Section 5.5 and 5.6 of this
task report, and recommendations are made in Section 5.7.
5.2 REVIEW OF REQUIREMENTS
Paragraph 85.478-15 (b) of EPA's Proposed Rulemaking
(NPRM) for New Motorcycles specifies the incorporation of a
variable speed blower system for simulating engine cooling.
The NPRM and the contract require that the blower have a
mechanism, controlled by the dynamometer roll speed, which
will regulate the blower flow to within 10 percent of the
roll speed over an operating range of 10 kph (6.2 mph) to
100 kph (62 mph). At speeds less than 10 kph, the air flow
shall be within ±1 kph of the simulated vehicle speed. The
5-1
-------
NPRM and the contract also call out a minimum duct exit area
2
of 0.5m and describe the relative positioning of the blower
exit with respect to the motorcycle. The effects the mileage
accumulation requirements would have on the design of the
system are discussed in this report.
5.3 GENERAL SPECIFICATIONS
In light of the requirements for emissions tests
as proposed in the NPRM and the contract, a basic set of
minimum specifications can be established. These are listed
in Table 5-1 and have been utilized to obtain from vendors
design data on suitable blower systems. Systems which do
not satisfy these specifications are also examined and
consideration is given to the effects their operation might
have on overall system performance and total cost.
Specifications for a blower system for the mileage
accumulation tests would be identical, except the maximum
linear velocity would increase to 30.56 m/sec (100.3 ft/sec)
and the corresponding blower capacity would be 916 m /min
(32,360 cfm).
5.4 DESCRIPTION OF EXISTING DYNAMOMETER/BLOWER SYSTEMS
A letter of inquiry was sent to all manufacturers
of vehicle and motorcycle chassis dynamometers requesting
literature on any motorcycle chassis dynamometers which they
might produce. In that inquiry and subsequent follow-up
conversations, information was obtained regarding engine
cooling systems which are currently in use. The design of
these systems vary. Most are constant-speed systems.
Several are multi-speed systems, and a few can truly be
classified as variable flow systems.
5-2
-------
Table 5-1. GENERAL SPECIFICATIONS
Minimum linear velocity at
blower outlet:
Maximum linear velocity at
blower outlet:
Minimum blower outlet area:
Maximum rate of change of
linear velocity:
Blower capacity (assuming ?
blower outlet area of 0.5 m ):
Static pressure at duct exit:
Noise level at maximum flow:
Response time of control system:
Degree of blower control (vehicle
speeds >10 kph):
Degree of blower control (vehicle
speeds <10 kph):
0.0 m/sec (0.0 ft/sec)
27.78 m/sec (9.1 ft/sec)
0.5 m2 (5.38 ft2)
1.47 m/sec2 (4.82 ft/sec2)
835 m3/min (29,500 cfm)
0.00 Pa. (0.00 in. H20)
85 dBA
1 sec.
±10% of vehicle speed
±0.3 m/sec (0.9 ft/sec)
5-3
-------
Burke E. Porter Machinery Company of Grand Rapids,
has recently completed the installation of three motorcycle
dynamometers in Kawasaki's new assembly plant in Lincoln,
Nebraska. These units, designed for short period production
3
line testing, are equipped with a 10 horsepower, 170 m /min
(6,000 cfm), centrifugal blower. This single-speed system
directs two jets of air towards the motorcycle and is designed
to provide sufficient cooling for full power and high-speed
testing of motorcycle engines up to 900 cc.
Hartzell Special Products, St. Paul, Minnesota,
one of the larger manufacturers of motorcycle dynamometers,
equips their units with single or dual blowers. Each blower
o
has a capacity of 11.3 m /min (400 cfm) and an exit area
2 2
less than 400 cm (160 in ). In the dual blower configura-
tion, the cooling jet is directed tangentially across the
side faces of the engine, while the jet from a single blower
installation is directed at an angle across the engine. In
both configurations, the exit of the blower is located just
ahead of the front wheel.
The Super Dyno built by J & R Manufacturing, Bell
Gardens, California, uses a small 3/4-horsepower blower,
operating at a constant speed. In their design the blower
is portable, permitting it to be placed in any position
relative to the motorcycle. PABATCO, Athena, Oregon, also
distributes a motorcycle dynamometer with a portable single-
speed blower system.
Carl Schenck Maschinenfabrick, represented in the
United States by Ostradyne Inc., produces two dynamometers
for motorcycle testing. One is a sprocket unit while the
other employs cradle rolls. These units, which are not
factory equipped with a blower, require the user to install
a blower which meets his respective requirements. Webco
Santa Monica, California, a manufacturer of motorcycle
parts, utilizes a two-speed blower with their Schenck dyna-
mometer, and speed selection is controlled by a temperature
5-4
-------
switch, controlled by the engine head temperature. Kawasaki,
also using a Schenck sprocket unit, uses a variable direction
blower. As the engine speed changes, adjustments are made
to the direction of the blower's guide vanes, changing the
total fraction of air which passes the engine.
Akkerman Engineering and Manufacturing, Houston,
manufactures a chassis dynamometer with a variable speed
blower system. Their system employs a (14,000 cfm) fan,
coupled to the hydraulic power absorber. As hydraulic
pressure is being generated by the hydraulic pump, most of
the pressurized fluid goes through a hydraulic motor to
operate the blower. As the motorcycle speed increases,
higher fluid pressures are produced, forcing more fluid
through the fan's hydraulic motor, and increasing fan speed.
2
The exit area of the blower is in excess of 0.5m and is
positioned directly in front of the motorcycle. However, no
attempt is made to equate the cooling air speed and the
speed of the motorcycle.
5.5 BLOWER CHARACTERISTICS
There are several types of blowers which might be
suitable for incorporation into the variable flow blower
cooling system. Each of these types will be considered
separately, and a number of the following blower design/
operational factors will be examined for each class of
blowers. These factors include:
t Type of fan
• Housing Characteristics
• Impeller Characteristics
• Discharge Air Pattern
• Efficiency
5-5
-------
• Air Flowrate vs. Pressure Drop
t Noise Level at Maximum Output
t Size/Configuration
• Duct Work Requirements
• Cost
• Delivery
In the following pages, the requirements for the
emissions certification test, as they relate to engine
cooling needs, will be discussed. The requirements for the
mileage accumulation test will be similar; only the magnitude
of the maximum airflow will be different.
5.5.1 Terminology
The following terms and symbols are used during
the discussion of the blowers, and control systems.
Total Pressure (TP) - the air pressure resulting
from passage of air
through the blower
Static Pressure (SP) - the total pressure rise
reduced by the velocity
pressure in the fan
Brake Horsepower (BMP) - the horsepower imparted
to the blower via the
shaft of the fan wheel
Air Horsepower - fan power output deter-
mined from the product of
air volume and total
pressure
5-6
-------
Mechanical Efficiency (ME) - the ratio of air
horsepower to brake
horsepower
Static Efficiency - the ratio between static
pressure and the horse-
power input
5.5.2 Centrifugal Blowers
There are four types of centrifugal blowers, each
utilizing a different impeller design. The four impeller
designs use:
• Airfoil blades
• Backward inclined/backward curved blades
• Radial and curved radial blades
• Forward curved blades
Each of these types has a different set of performance
characteristics and noise emissions, and each has been
designed for different applications. Design and performance
data for all types of centrifugal blowers have been solicited
from manufacturers. Data on centrifugal units which meet
the specifications outlined in Table 5-1 and the above
criteria have been tabulated in Table 5-2.
The characteristics of this generic class of
blowers are described below in Section 5.5.2.1. Specific
characteristics resulting from the impeller design are
described separately.
5.5.2.1 General Description
In this class of blowers, the air pressure results
from the centrifugal force created by rotating the air
5-7
-------
CHARACTERISTICS OF CENTRIFUGAL BLOWERS
MANUFACTURER
Aladdin Heating
Chicago Blower
Barry Blower
New York Blower
IMPELLER
TYPE
Bckwrd. Incl.
Radial
Airfoil
Bckwrd. Incl.
Airfoil
Bckwrd. Incl.
Curved Radial
Radial
MODEL
50BR
57BR
64BR
SOAR
57AR
64AK
36-1/2 SISW
40-1/4 SISW
44-1/2 SISW
30 DIDW
33 DIDH
365 SISW
402 SISW
445 SISW
270 DIDW
3OO DIBW
449 SISW
339 DJDW
448
784 DH
784 LS
INLET
DIA.
(m)
0.737
0.838
0.940
0.737
0.838
O.94O
_
_
_
_
-
_
_
_
_
-
1.130
0.838
1.130
1.143
1.143
OUTLET
HT
(nil
0.714
O.B13
0.908
0.714
0.813
0.908
0.398
0.997
1.102
0.708
0.784
0.983
1.081
1.198
0.727
0.808
1.260
_
1.260
1.127
1.127
SIZE
WIDTH
In)
0.619
0.702
0,787
0.619
0.702
0.787
0.8O6
0.895
0.990
1.216
1.338
0.733
0.814
0.395
0.973
1.078
0.851
_
0.851
0.937
0.937
OUTLET
AREA
In )
0.442
0.570
0.715
0.442
0.570
0.715
0.724
0.893
1.092
0.861
1.049
0.720
0.880
1.072
0.707
0.811
1.072
1.047
1.072
1.056
i.056
AIR
FLOW
(«i /mini
856
840
846
856
840
846
839
859
853
840
832
868
844
839
854
844
839
830
839
880
880
SP - .75 kPa (V H20)
RPM
788
587
4S5
185
585
469
1159
929
73 B
1245
1000
1196
930
747
1574
1224
B25
1100
757
333
336
BfH
46.7
33.8
29.4
4fi.7
36.4
29. f
21.8
24.1
20.5
23.0
19.8
31.6
24.6
20.9
31.4
24.9
18.6
19.6
20.0
23.?
26.4
SP - 1.0 kPa (4" H20)
RPM
831
641
527
828
628
510
1206
979
793
1314
1070
1248
986
806
1649
1304
884
1184
820
370
375
blip
52.7
40.5
35.2
57.2
42.1
35.6
32.2
28.7
25.4
21 . 7
24.4
36.2
29.3
25.8
36.1
29. 8
23.4
24.8
25.2
23.6
33.5
EXPECTED
SP
(k-Pal
0.70
0.70
0.70
0.70
0.70
0.70
0.75
0.80
0.85
0.95
1.00
0.70
0.75
0.80
0.85
0.90
0.80
0.80
0.80
0.80
0.80
MASS
(kgl
1202
1515
1894
1202
1515
1894
-
-
-
-
-
450
599
748
319
444
974
513
975
3513
3314
APPROX. SIZE
HEIGHT
(m)
2.080
2.356
2.626
2.080
2.356
2,626
1.810
2.007
2.210
1.638
1.810
1.745
1.932
2.129
1.307
1.448
-
-
-
3.400
3.400
WIDTH
(n)
1.810
2.089
2.286
1.810
2.089
2.286
1.608
1.637
1.775
1.591
1.759
1.073
1.168
1.273
1.340
1.403
-
-
-
2.505
2.505
DEPTH
(ml
1.820
2.064
2.288
1.820
2.064
2.288
1.633
1.784
1.964
1.334
1.465
1.537
1.694
1.832
1.135
1.265
-
-
-
3.245
3.245
PRICE t S )
F.O.B. MFC
F.O.B. Olsor
_
-
-
-
-
-
1622
2124
2469
1756
3770
-
-
3170*
-
3330*
2207
1740
2220
7265
6605
NOTES
*
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Belt Drive
Direct or Belt
Direct or Belt
Direct or Belt
Direct or Belt
Direct or Belt
Direct or Belt
Direct or Belt
Direct or Belt
Belt Drive
Belt Drive
01
I
OS
-------
column between the blades and by the kinetic energy imparted
to the air by virtue of its velocity leaving the impeller.
The air velocity resulting from the rotative velocity is a
result of the impeller design and the fan flow rate.
The majority of centrifugal blowers have single
inlets (SISW), but some units utilize inlets on both sides
of the blower. These latter units, denoted as DIDW units,
are typically smaller than SISW units having similar character-
istics. As an example, Chicago Blower's Model 40% SISW and
Model 30 DIDW are similar units. The outlet areas of the
two units are almost identical, but the DIDW is smaller in
size, operates at a higher impeller speed, and requires
slightly less power. Similar comparisons can be made between
the DIDW and SISW units fabricated by Barry Blower and New
York Blower.
The centrifugal blower is driven directly by a
motor coupled to the impeller shaft or indirectly by a motor
connected to the shaft by a belt. The selection of direct-
coupled motors or belt-driven motors is not always available
to the user since many blower mounting configurations do not
permit the use of direct-drive motors. When belt-drive
motors are used, some slippage between the belt and shaft
occurs, reducing the efficiency of the system. As a result
of the frequent high acceleration and deceleration torque
requirements, this slippage would be unacceptable if a
variable-speed motor were used to attain the variable flow
control. Therefore, in order to obtain maximum efficiency
and permit the usage of variable-speed motor control systems,
only blower mounting arrangements which permit the use of
direct-drive motors will be considered.
The fan casing, commonly called the scroll or
spiral, collects the air delivered from the impeller and
directs the flow into the exit duct. As a result of the
curvature of the scroll and the centrifugal action of the
impeller, the resultant velocity profile at the exit of the
5-9
-------
blower is nonuniform. In most standard applications, this
nonuniformity is not a problem, but it may not be acceptable
for this application. Directing vanes can be used to
transition the spiral flow in the housing to streamline flow
in the exit duct. These vanes can also be used to restructure
the velocity profile, producing a more uniform profile.
Typically, the exit configuration of centrifugal
fans is rectangular in shape and a scroll configuration can
be selected that would result in the outlet duct being
positioned just above and parallel to the floor. However,
even though the exit duct is appropriately located with
respect to the motorcycle, some additional ducting is required,
This will include telescoping ducting in order to position
the duct exit just in front of the front wheel of the motor-
cycle; flow straightening and conditioning sections; and a
transition section from the blower exit duct dimensions to
p 2
the required exit area of at least 0.5m (5.4 feet ).
The pressure loss at maximum flow in this ducting
can be approximated by the following equation:
Hd (kPa) =• 0.45 + 0.45 C1 + 0.03n + H$ (5-1)
The first term accounts for the pressure drop which occurs
at the exit of the duct when the velocity pressure of the
cooling air is converted to static pressure. The second
term describes the pressure drop associated with a change of
duct area from the exit area of the blower duct to the final
2
exit area of at least 0.5m . The coefficient C,, is a
function of the ratio of area of the exit blower duct to the
2
final exit area of*0.5m . The magnitude of C, will also
depend on whether the transition is gradual or abrupt. The
next term accounts for the losses attributable to the straight-
ening vanes, where n is the number of vanes. The final term
HS is the pressure loss across a noise baffle, if it should
be used. Considering all of these terms, it is anticipated
5-10
-------
that the total pressure drop would be 0.50 - 0.88 kPa (2 -
3% in. H20). Approximate values for the pressure drop for
the various blowers considered are tabulated in Table 5-2.
Should a centrifugal blower be used for this
application, it would have to be firmly mounted to the floor
of the test cell. This would be necessary because of the
mass of the blower, which, excluding motor and controls,
typically weighs more than 900 kg (2,000 Ibs), and the
effect vibrations might have on the mechanical balance of
the impeller wheel and, thus, the performance of the blower
system. The blower would be mounted so that the exit duct
would be centered about the front wheel restraints of the
dynamometer, thereby providing symmetric cooling to the
motorcycle engine. A similar mounting scheme could be used
on a system for testing three-wheeled motorcycles.
The fan wheel inertia is also an important para-
meter, particularly when a variable-speed motor control is
used. The relationship between motor size and the wheel
inertia is discussed in Section 6. In centrifugal blowers
the magnitude of the fan wheel is substantial. The inertia
of New York Blower's Size 449 and 448 units is just under
25 kgm2 (600 Ib ft2) while equivalent DIDW units are approxi-
mately one-half that size.
As shown in Figures 5-1 to 5-4, the sound power
levels from a blower operating at the anticipated conditions
leave the fan wheel at a velocity slower than the impeller
tip speed. Also as a result of the blade depth, efficient
are substantial. Any reductions in these levels would
reduce the requirements for sound insulation and provide a
better working environment for the vehicle driver. Substan-
tial noise reductions can be obtained by utilizing a fan
silencer. These devices are commercially available and can
be attached directly to the inlet or outlet of the blower.
The reduction in sound power levels is shown in Table 5-3.
This magnitude of reduction can be achieved in all types of
blowers, not just centrifugal units.
5-11
-------
IMPELLER CHARACTERISTIC
BLOWER CHARACTERISTICS
o
. z
UJ LU
OO
•
.
O
0.
U. UJ
o co
cd
I— O
(_) UJ
o: v:
LU <
a. or
CD
BHP
PERCENT OF WIDE OPEN VOLUME
SPECIFIC SOUND POWER LEVEL
dB re 10" watt, 849m3/min G> 1 kPa Fan Total Pressure
OCTAVE
CENTER FREO
LEVEL
1
63
92
2
125
92
3
250
91
I)
500
89
5
1 000
88
6
2000
83
7
^ 000
75
8
8000
67
Figure 5-1. CENTRIFUGAL BLOWER AIRFOIL IMPELLER
5-12
-------
IMPELLER CONFIGURAT ION
ce.
o
•
.
.
CO
•
:
:
BLOWER CHARACTERISTICS
HP
0 10 20 30 40 50 60 70 80
PERCENT OF WIDE OPEN VOLUME
00
o
&•-
SPECIFIC SOUND POWER LEVEL
dB re 10"12 watt, 849m3/min @ 1 kPa Fan Total Pressure
OCTAVE
CENTER FREO
LEVEL
1
63
92
2
125
92
3
250
91
k
500
89
5
1000
88
6
2000
83
7
4000
75
8
8000
67
Figure 5-2. CENTRIFUGAL BLOWER
BACKWARD INCLINED/BACKWARD CURVED IMPELLER
5-13
-------
MPELLER CONFIGURATION
•
.
• >
•
LU
)
GO
UJ
X
00
40 -
20
BLOWER CHARACTERISTICS
: HP
TP
ME
SP
SE
l
20 *iO 60 80
PERCENT OF WIDE OPEN VOLUME
00
C_J
SPECIFIC SOUND POWER LEVEL
dB re 10" watt, 849m3/min @ 1 kPa Fan Total Pressure
OCTAVE
CENTER FREQ
LEVEL
1
63
105
2
125
102
3
250
1 00
k
500
1 00
5
1000
95
6
2000
90
7
iiOOO
87
8
8000
86
Figure 5-3. CENTRIFUGAL BLOWER
RADIAL IMPELLER
5-14
-------
MPELLER CONFIGURATION
...
00
LU
C£.
o.
BLOWER CHARACTERISTICS
20 kO 60 80
PERCENT OF WIDE OPEN VOLUME
:HP
TP
cSP
f SE
UJ
.
SPECIFIC SOUND POWER LEVEL
dB re 10 watt, 849m3/min @ 1 kPa Fan Total Pressure
OCTAVE
CENTER FREO
LEVEL
1
63
97
2
125
95
3
250
95
I|
500
91
5
1 000
85
6
2000
81
7
ifOOO
78
8
8000
72
Figure 5-4.
CENTRIFUGAL BLOWER
FORWARD CURVED IMPELLER
5-15
-------
Table 5-3. TYPICAL SOUND POWER LEVEL
REDUCTIONS BY FAN SILENCERS
Octave
Center
Dynamic
Insert!
Loss (d
Band
Frequency
on
B)
Veloc
0
100
300
Ity
m/sec
m/sec
1
63
3
2
2
2
125
6
5
4
3
250
16
15
14
4
500
23
22
21
5
1000
25
24
23
6
2000
23
22
21
7
4000
18
17
16
8
8000
16
15
14
There are several penalties associated with the
installation of a fan silencer. The most obvious penalty is
the cost of the unit. Also, as a result of installing the
silencer on the blower outlet, a substantial pressure loss
is incurred. The magnitude of that loss can almost double
the total system pressure drop and increase the horsepower
requirements by as much as 20 percent. However, since the
silencer may also serve as a mixing baffle, its installation
may improve the uniformity of the air exiting from the duct.
5.5.2.2 Airfoil Impellers
In this design, the impeller, which is shown in
Figure 5-1, consists of ten to sixteen airfoil blades curved
away from the direction of rotation. This causes the air to
leave the fan wheel at a velocity slower than the impeller
tip speed. Also as a result of the blade depth, efficient
expansion of the air occurs within the blade passages.
These features result in a blower design which is highly
efficient and relatively quiet.
The performance characteristics are also shown in
Figure 5-1. While the sound power levels are high, they are
lowest of all the fans examined in this report. The perform-
ance curves show that the peak efficiency occurs when the
fan is operating at 50 to 60 percent of free flow. The peak
efficiency of 80 to 85 percent is higher than that obtained
by other types of blowers. This also is an operating area
5-16
-------
where the pressure curve is rising, thus providing good
operational stability near the point of peak power consump-
tion. The static pressure trace has been normalized so that
the 100 percent point is at the point of maximum efficiency.
As seen in Table 5-2, the airfoil designs, built
by New York Blower and Chicago Blower, require less power to
reach the desired flow condition than do the blowers equip-
ped with backward inclined or radial blades. However, in
order to take advantage of this design's inherent efficien-
cies, it is necessary to go to a larger scroll and fan wheel
size. This results in a larger blower system, and substant-
ially larger exit duct areas which, in turn, result in a
larger pressure drop when transitioning down to the exit
2 2
duct area of 0.5m (5.4 ft ). Relative costs for these
blowers are also shown in Table 5-2.
5.5.2.3 Backward Curved/Contoured Blades
This impeller design, which is sketched in Figure 5-2,
is very similar to the airfoil design. The only real differ-
ence is that the blades are of uniform thickness and back-
wardly curved or inclined. As seen in the performance
curves, this results in a slight reduction in efficiency,
1 to 5 percent, and a negligible increase in sound output,
as compared to the airfoil design.
As shown in Table 5-2, several blowers of this
type are available which would only require a minimal exit
duct transition piece. However, as seen in the specifica-
tions of the Aladdin units (Table 5-2), a severe horsepower
penalty is incurred by going to the unit with the smaller
exit area. The smaller unit is slightly less expensive, but
the costs for the smaller motor required by the larger unit
may offset this price differential, and result in a lower
net price for the larger system.
5-17
-------
5.5.2.4 Radial Blades
In this design, six to ten blades are radially
mounted on the impeller. The blades may be as shown in
Figure 5-3, or they may be slightly curved. The resultant
design has high mechanical strength and is easily repaired.
It has been extensively used in material handling applica-
tions where equipment ruggedness and design simplicity are
essential. The size, weight, and cost of these blowers is
substantially greater than airfoil or backwardly inclined
blowers having comparable characteristics. The data on the
New York Blowers, tabulated in Table 5-2, shows the
relationships between the three styles of blowers.
As shown in the performance curves in Figure 5-3,
this style of blower has substantially higher pressure
characteristics than the other types of blowers, but the
radial type has a lower efficiency throughout its entire
range of operation. The sound power output levels for this
style are also the highest of all the blowers being considered
In contrast to the previously described centrifugal blowers,
the horsepower curve for radial blowers shows a continual
rise to a maximum which occurs at free flow.
5.5.2.5 Forward Curved Blades
Blowers equipped with forward curved blades are
used primarily in applications such as residential heating
and ventilation units and packaged air conditioning equipment
which requires low pressure and small to moderate air volumes.
The performance curves and sound characteristics of this
style are displayed in Figure 5-4.
Vendors did not supply any data on this type of
blower, because it is incapable of meeting this application's
air volume requirements. Therefore, further consideration
of this style is unwarranted.
5-18
-------
5.5.3 Axial Blowers
Axial blowers are characterized by the mechanism
in which they produce velocity and pressure changes. In
this class, air pressure is produced only as a result of
forward velocity changes which occur when air passes through
the fan wheel.
Three types of blowers can be classified as axial
units. These are:
t Propellers
t Tubeaxial Fans
• Vaneaxial Fans
The three types of axial units have similar design
and performance characteristics which are discussed below in
Section 5.5.3.1. The characteristics specific to each style
are considered separately.
Data on axial units, which meet the requirements
outlined in Table 5-1, have been solicited from vendors, and
have been tabulated in Table 5-4.
5.5.3.1 General Description
In contrast to centrifugal blowers, axial blowers
have both a circular inlet and outlet, usually of comparable
areas. In many applications, an entrance orifice is installed
at the blower inlet. This curved inlet addition reduces the
entrance pressure loss on installations where no duct work
precedes the fan. Since the outlet of the axial blower is
circular, ducting will be required to transition the air
from this circular cross section to the rectangular configur-
ation required at the exit. However, the pressure drop
which results in this geometry transition is no greater than
that incurred in the reduction of a centrifugal blower exit
area from one rectangular size to another.
5-19
-------
Table 5-4.
CHARACTERISTICS OF AXIAL AND MIXED AXIAL BLOWERS
MANUFACTURERS
Aero vent
Hartzell
Pacific Air Industries
TYPE
Vaneaxial
Vaneaxial
Mixed Axial
MODEL
V311-Y42
V361-Y34
V361-Y38
V361-Y42
V381-Y30
VA36A
50 36VR3
270
INLET
DIA.
(m)
0.967
1.023
1.023
1.023
1.094
1.041
0.914
0.944
OUTLET
DIA.
(m)
0.846
0.916
0.916
0.916
0.973
0.914
0.914
0.944
OUTLET
AREA
(m )
0.562
0.659
0.659
0.659
0.744
0.656
0.656
0.700
RPM
1750
1750
1750
1750
1750
1750
1750
1750
SP = 0.75 kPa
AIR FLOW
(m /mini
714
814
903
980
854
823
863
792
(3" HO)
BHP
21.1
23.7
29.7
34.2
22.9
24.7
25.5
42.0
SP = 1 kPa
AIB FLOW
(m /min)
568
731
816
888
753
-
807
-
(4" HO)
BHP
22.9
25.0
30.4
34.7
24.1
-
28.5
-
EXPECTED
SP
(kPa)
0.60
0.65
0.65
0.65
0.65
-
0.65
0.65
MASS
(kg)
324
395
426
474
442
-
-
-
APPROX. SIZE
MAX blA.
(m)
0.967
1.023
1.023
1.023
1.094
1.041
1.029
1.048
LENGTH
(m)
0.762
0.838
0.838
0.838
0.838
1.371
1.524
1.303
COST
(S)
-
-
-
-
-
657
987
2010
NOTES
Direct Drive
Direct Drive
Direct Drive
Direct Drive
Direct Drive
Direct Drive
Direct Drive
Belt Drive
CJ1
I
ro
o
-------
The pressure loss in systems using this class of
fan can also be estimated by an equation identical to
equation (5-1).
Physically, axial units are smaller and weigh less
than comparable centrifugal units. In many cases, the
weight difference is greater than 50 percent. As a result,
the axial blower may be able to be mounted in such a manner
that the entire blower can be repositioned so as to maintain
a constant spacing between the exit of the air duct and the
front wheel of the motorcycle. This would be an alternative
to mounting the blower firmly to the floor of the test cell
and utilizing telescoping ducts to maintain the required
separation.
Axial blowers can also be operated by a motor
directly attached to the fan or indirectly by a belt-driven
system. Again, direct-drive systems are more efficient than
the belt-driven system because slippages of the belt reduce
the systems' efficiency. As concluded in the discussion of
the centrifugal units, only axial units which permit the use
of direct-drive motors should be considered because of the
higher efficiency and adaptability to variable speed motors.
Fan silencers can also be used in conjunction with
axial blowers, and sound reductions comparable to those
shown in Table 5-3 can be achieved. However, most axial
units would not be able to overcome the pressure drop intro-
duced by the silencer. Flow straighteners could be used in
conjunction with these units, but they would not appreciably
reduce the noise emitted by these units.
5.5.3.2 Propeller
The propeller is the simplest axial fan design and
consists of two or more single thickness blades attached to
a relatively small hub. This design, which is shown in
Figure 5-5, is capable of creating high flow rates, but only
5-21
-------
PROPELLER CONFIGURATION
!
I
-
o
>-
;
-
X
•a:
100
80
60
20
BLOWER CHARACTERISTICS
0 20 40 60 80
PERCENT OF WIDE OPEN VOLUME
00
ME
SE
HP
TP
SP
<
SPECIFIC SOUND POWER LEVEL
dB re 10" watt, 849m3/min @ 1 kPa Fan Total Pressure
OCTAVE
CENTER FREQ
LEVEL
1
63
108
2
125
105
3
250
106
4
500
104
5
1000
102
6
2000
102
7
4000
100
8
8000
88
Figure 5-5. AXIAL BLOWER
PROPELLER
5-22
-------
at low pressure levels. As seen in the performance curves,
the efficiency of this style is very poor, and the sound
power levels are very high. The discharge pattern of the
air is circular in shape, and the airstream swirls as a
result of the action of the blades and the lack of straight-
ening vanes.
No propellers were identified which would be
suitable for this application. Those which could meet the
flow requirements would not be able to overcome the pressure
drop which would be produced in the outlet transition duct
and flow straightening section.
5.5.3.3 Tubeaxial Fans
As shown in Figure 5-6, a tubeaxial fan consists
of a propeller or fan wheel mounted in a cylindrical tube.
Clearance between the wheel tip and the inner wall of the
tube is very small. The propeller fan typically consists of
four to eight, single thickness or airfoil blades attached
to the wheel hub. The blade pitch affects the blower
characteristics.
As a result of placing the propeller within the
tube, improved efficiency at low and moderate pressure
levels is obtained. This can be seen in the performance
curves shown in Figure 5-6. These curves also show a dip in
the pressure curve just to the left of the peak pressure
point. This dip is caused by aerodynamic stall, and operation
at this condition should be avoided. Resultant sound power
levels are significantly lower than those for propellers,
but higher than those for centrifugal blowers'. The discharge
air pattern from these units also exhibits a rotational or
whirling motion which is a result of the rotation of the
propeller and would require straightening vanes.
No information on suitable tubeaxial units was
supplied by vendors.
5-23
-------
TUBEAXIAL CONFIGURATION
.
C:
-
X
• :
BLOWER CHARACTERISTICS
20 kO 60 80 100
PERCENT OF WIDE OPEN VOLUME
.
SPECIFIC SOUND POWER LEVEL
dB re 10" watt, 849m3/min @ 1 kPa FAN TOTAL PRESSURE
OCTAVE
CENTER FREO
LEVEL
1
63
101
2
125
99
3
250
103
k
500
101
5
1000
99
6
2000
97
7
4000
9^
8
8000
87
Figure 5-6. AXIAL BLOWER
TUBEAXIAL
5-24
-------
5.5.3.4 Vaneaxial Fans
Vaneaxial fans are similar to tubeaxial units.
The major difference is the inclusion of guide vanes just
downstream from the fan wheel. The guide vanes straighten
out the rotary motion imparted to the air by the rotation of
the wheel. This improves the efficiency and pressure charac-
teristics of the fan, as shown in Figure 5-7.
In order to achieve further improvement in these
characteristics, airfoil blades are used and attached to a
wheel hub which has a diameter greater than 50 percent of
the wheel tip diameter. This further reduces the sound
power levels, as shown in Figure 5-7.
The pitch of the blade is also an important charac-
teristic as can be seen from the data tabulated in Table 5-5.
The effects of pitch angle can be seen from the data on
Aerovent's Series V 361 fans. The data indicates that by
increasing the pitch from 86 cm (34 inches) to 107 cm (42
inches), the volumetric air flow at 0.50 kPa (2 inches HpO)
static pressure can be increased by almost 25 percent with a
comparable increase in horsepower, but with no increase in
the rotational velocity of the unit.
The rotational inertia of a vaneaxial fan wheel is
less than l/10th of that a centrifugal fan wheel having
comparable performance. Therefore, although these units
operate at higher wheel speeds per volume air flow delivered
than centrifugal units, they will have higher acceleration
rates and will require less power to accelerate the wheel.
In contrast.to propellers or tubeaxial units,
vaneaxial units are a viable means for satisfying the vari-
able flow cooling problems of this application.
5-25
-------
VANEAXIAL CONFIGURATION
•
Q.
LU
CO
o
••
,
• r
0
BLOWER CHARACTERISTICS
20 ifO 60 80
PERCENT OF WIDE OPEN VOLUME
11
: :
:
SPECIFIC SOUND POWER LEVEl
dB re 10'12 watt, 849m3/min @ 1 kPa FAN TOTAL PRESSURE
OCTAVE
CENTER FREQ
LEVEL
1
63
99
2
125
96
3
250
98
1»
500
99
5
1000
97
6
2000
9^
7
4000
92
8
8000
82
Figure 5-7.
AXIAL BLOWER
VANEAXIAL
5-26
-------
5.5.4 Mixed Axial Centrifugal Blowers
The mixed axial centrifugal blower combines the
space saving features of the axial fan, i.e., 40 percent
less installation space than centrifugal units, with the
centrifugal blowers favorable performance characteristics,
such as lower operating speeds and lower noise levels than
associated with axial units. This is achieved by a unique
air flow design which is shown in Figure 5-8. The air is
drawn from the inlet into a centrifugal, backwardly inclined
impeller and is discharged radially from the fan wheel. The
air then must change direction by 90° where it enters the
guide vane section.
As a result, as seen in Figure 5-8, this fan's
efficiency curve is comparable to a vaneaxial's, but slightly
lower than the curve of a backwardly inclined centrifugal
fan. Noise emissions are moderate, somewhat greater than a
quiet centrifugal but lower than a vaneaxial unit.
The principal difference can be found in the
pressure curves. The mixed axial's pressure curve shows a
continually rising pressure curve characteristic, which
permits the fan to operate throughout the pressure curve
range without overload. Centrifugal or axial fans with
peaking or reversing pressure curve characteristics cannot
be operated to the left of the pressure peak because of air
pulsations, excessive noise, and fan vibration.
Data on a suitable mixed axial centrifugal fan has
been tabulated in Table 5-4.
5.6 VARIABLE FLOW CONTROL METHOD CHARACTERISTICS
The examination of a flow control system has been
separated from the study of suitable blowers and fans because
the control methodology is not dependent on the blower
5-27
-------
MIXED AXIAL CENTRIFUGAL CONFIGURATION
•
BLOWER CHARACTERIST ICS
20 ^0 60 80
PERCENT OF WIDE OPEN VOLUME
00
SPECIFIC SOUND POWER LEVEL
dB re 10"12 watt, 849m3/min @ 1 kPa FAN TOTAL PRESSURE
OCTAVE
CENTER FREQ
LEVEL
1
63
103
2
125
1 00
3
250
100
k
500
95
5
1000
3k
6
2000
89
7
^000
85
8
8000
82
Figure 5-8. MIXED AXIAL CENTRIFUGAL BLOWER
5-28
-------
design. Only when the control system is interfaced with the
blower must the design parameters of the blower be explicity
defined. In addition, no commercially available variable
flow blower system has been identified. Rather, these
systems are designed by the user to meet the specific
requirements of the application.
As stated previously, a control system for the
blower is required to equate the air speed at the exit of
the blower ducting with the dynamometer roll speed. The
control system should be capable of varying the blower flow
rate over at least a 100-to-l range. This large a range is
required in order to meet the maximum flow rate requirement
of 100 km/hr and the accuracy requirements of ±1 km/hr at
dynamometer speeds less than 10 km/hr as specified in the
control. An estimated response time of less than 1/4
second would be required to maintain the equivalence of air
speed and vehicle speed within the accuracy limits specified
in the contract. The exact degree of control and range of
variable flow can best be specified after quantitative data
have been obtained on the relationship of engine temperature
and motorcycle exhaust emissions. Therefore, systems which
meet either less stringent or more demanding specifications
are examined in this report.,
The discussion of the variable flow control methods
has been divided up into two parts. In the first, the
suitability of both open-ended and closed-looped systems
will be examined. The other part of the discussion will
closely examine the specific methods of achieving a variable
flow control system for a blower.
5.6.1 Control Methods
Either an open-ended or closed-loop system can be
utilized to control the variable flow blower system.
5-29
-------
The open-ended system is shown schematically in
Figure 5-9. In this system a change in motorcycle speed is
sensed by the dynamometer tachometer. This signal is used
to directly alter a precalibrated mechanism which modifies
the blower flow. Since the linkage between the speed signal
and the control mechanism is "hard," the response time is
minimal and depends only on the electronic response time of
the amplifiers and any control mechanism actuators.
The inherent simplicity of the open loop system is
attractive. However, a mechanical or electrical interface
is required unless the control mechanism is inherently
linear. For most applications either mechanical or electrical
linearization devices can be designed, but extensive design
and testing is usually required to "time" the linearizer and
establish system reliability because of the uniqueness of
each system. The performance of the open-ended system will
depend on the accuracy of any linearizer and on the repeata-
bility and stability of the controller's cause and effect
relationship.
In a closed loop system a feedback sensor is
added, as shown in Figure 5-10. The closed-loop control
system does not require linearization or calibration of the
blower control mechanism, but linearization of the signal
from the feedback sensor may be required. The selection of
components for the feedback system is critical because their
performance characteristics define the stability and response
time of the control system. Specifically, if the feedback
sensor is a tachometer mounted on a fan wheel shaft, changes
in shaft speed will sensed almost instantaneously and there
will be a negligible effect on the system response time. If
a velocity sensor is used, the time required for the air
column to reach the sensor, in addition to the sensor response
time, may define the system response. Therefore, consider-
ation should be given to the characteristics of available
velocity sensors. The techniques which are available include:
5-30
-------
TACH
SIGNAL
G
BLOWER
OUTPUT
CONTROL
Figure 5-9. OPEN-ENDED CONTROL SYSTEM
TACH SIGNAL
O
CONTROL
H
FEEDBACK
BLOWER OUTPUT
Figure 5-10. CLOSED-LOOP CONTROL SYSTEM
5-31
-------
• Pitot tubes
• Turbine meters
• Heated thermopiles
• Hot wire anemometry
The pitot tube approach is extensively used in air
pollution monitoring and investigations of heating/ventilating
systems. It is accurate, usually within ±2 percent, and the
velocity determination in this system is a function of the
square root of pressure difference. Therefore, a lineariza-
tion circuit would have to be developed.
Turbine meters are not suitable for this application
because they cannot operate over the 100-to-l dynamic range
imposed by the contract specifications. Typically they
function over a 20-to-l range.
Heated thermopiles and hot wire anemometers utilize
similar technologies to sense velocity. Since the parameter
which is measured is the change of heat flux, the units tend
to have a fast response time, in the range of 100-to-l,000
milliseconds, and accuracy within ±2 percent over a dynamic
velocity range of 100-to-l. Inherently these sensors are
nonlinear, but several manufacturers have developed electronic
linearization circuits.
With any velocity sensor the positioning of the
sensor and the design of the control loop are critical for
minimization of instabilities and maximization of the control
accuracy. The use of velocity feedback sensors is examined
further in Section 6 of this report.
5.6.2 Control Mechanisms
Five blower control mechanisms which may be applic-
able are:
5-32
-------
• Dampers
§ Variable Speed DC Motor
• Constant Speed Motor with Eddy Current Clutch
t Variable Bypass Flow
• Compound Control System
Each of these mechanisms will be examined, and their advantages
and disadvantages detailed. A summary of the findings has
been incorporated into Table 5-5.
5.6.2.1 Shut-Off Dampers
Dampers can be placed either on the inlet to the
blower or at the exit. While the design of inlet and outlet
dampers differ, the effect is the same; they restrict the
amount of air a blower can deliver.
Outlet dampers are only used in conjunction with
centrifugal blowers and are available in two different
designs, parallel blade and opposed blade. The parallel
blade design, shown in Figure 5-11, consists of four to six
unidirectionally-oriented louvered vanes which effectively
restrict the flow from the blower. As seen in Figure 5-11,
the magnitude of the throttling action is a nonlinear function
of the damper setting. As a result of the nonlinearity,
this style is best suited for applications requiring accurate
air volume control over a narrow range of flow; for example,
from wide open to 75 percent of wide open capacity.
Opposed blade dampers, shown in Figure 5-12,
utilize vanes which are oppositely louvered. This results
in a slightly costlier design, but yields an almost linear
control mechanism, and is, therefore, suited for control
over a broad range of air volumes. However, operation at
low flow rates is restricted because of the inherent insta-
bility of centrifugal blowers operating at low fractions of
wide open capacity and the constant leakage through a
5-33
-------
Table 5-5. SUMMARY OF CONTROL SYSTEMS
CONTROL SYSTEM
(suppliers)
OUTLET DAMPER
(Blower Mfgs)
INLET DAMPERS
(Blower Mfgs)
BYPASS - SINGLE
DUCT
(Olson Labs)
BYPASS - DUAL
DUCT
(Olson Labs)
EDDY CURRENT
DRIVE
(Eaton Dynamic)
(Louis Allis)
DC DRIVE
(Eaton Dynamic)
(Louis Allis)
(Randtronics)
(Sabina Electric)
(Reliance Elec-
tric)
REGENERATIVE DC-
DRIVE
(Randtronics)
(Sabina Electric)
CONTROL
CLOSED
LOOP
X
X
X
X
METHOD
OPEN
LOOP
X
X
X
FEEDBACK
SENSOR
Veloc-
ity
—
—
-
Tach
Tach
Tach
REGU-
LATION
%
2%
2%
2%
2%
1%
0.5%
0.25%
AIR FLO
MIN.
(km/hr
15
15
0
0
3
1
0
W RANGE
MAX.
(km/hr)
100
100
100
100
100
100
100
APPROX
COST.
$1500
$1500
$4000
$4000
$6000
$6800
$8000
REMARKS
System must be de-
signed and tested.
System roust be de-
signed and tested.
Dynamic braking may
be required .
Dynamic braking may
be required.
in
i
-------
20 30 40 SO 60 70 10 90 100
PERCENT OF WIDE OPEN CAPACITY
Figure 5-11. PARALLEL BLADE OUTLET DAMPER
5-3b
-------
0
90*
10*
70'
60*
•" 40'
2 Jo*
ft.
< 20*
o
10»
0*
WIDE
OPEN
0 10 20 30 40 50 60 70 10 90 100
PERCENT OF WIDE OPEN CAPACITY
Figure 5-12. OPPOSED BLADE OUTLET DAMPER
5-36
-------
supposedly closed damper. This leakage can be as high as
20 percent of wide open capacity.
An inlet vane control mechanism, as shown in
Figure 5-13, is a nonlinear control device which is compatible
with both centrifugal and vaneaxial blowers. This style of
control provides more economical performance at lower air
volumes than the outlet damper. The power savings for a
centrifugal fan is graphically shown in Figure 5-14. The
reduction in horsepower is achieved by using the inlet vanes
to pre-spin the entering air in the same direction as the
wheel rotation. Reduced operating costs at partial loads
plus the use of the lowest cost, constant speed drive are
additional advantages. However, even inlet dampers display
significant leaks, approximately 15 percent of wide open
capacity, under "shut-off" conditions.
As shown in Figures 5-11 to 5-13, the flow through
the dampers is nonlinear with the degree of damper opening.
However, a suitable linkage system can be designed so as to
linearize this actuation. In addition, the damper vanes
should be mounted in ball bearings rather than sleeve-type
bearings to reduce friction and increase the positioning
accuracy of the vanes. The positioning of the vanes is
critical, and serious consideration should be given to the
use of a closed-loop control if repeatable positioning of
the dampers cannot be assured.
The leakage through either inlet or outlet dampers
can be almost eliminated by using special felt-tipped vane-
edge wiper seals. However when closed, the blower would be
forced to operate in a "near shut-off" condition. When
operated in this region, the blower, regardless of whether
it is a centrifugal or a vaneaxial unit, may stall, vibrate
and operate unstably. Therefore, as long as cooling air
flows of less than 15 km/hr, are required, a shut-off damper
control should not be considered. However, because of their
low cost, $500 for damper, $1,000 to 2,000 for control
5-37
-------
0 10 20 30 40 20 60 7Q 80 90 100
PERCENT OF WIDE OPEN CFM and BMP
Figure 5-13. INLET DAMPER
5-38
-------
60
Q.
O
t-
t/1
It
LiJ
O
Q-
UJ
CO
O
t
• i
Q
CO
O
120
00
80
60
20
FAN STAT 1C
PRESSURE
BHP OUTLET
DAMPER
CONTROL
OPERATING
POINT
r^BHP INLET-y
h VOLUME >£*
CONTROL/^
— ^» 1
\
-SYSTEM RESISTANCE \
1 I i-
\
i \
20 40 60 80 100 120
VOLUME - PERCENT OF RATING
40
Figure 5-14. COMPARISON CURVES
OF INLET AND OUTLET DAMPERS
5-39
-------
linkages and actuators, their use should be considered it
flows less than 15 km/hr are not required.
5.6.2.2. Bypass Damper Flow Control
Two concepts of the bypass control method are
shown schematically in Figure 5-15. In these configurations
the fan is operated at a constant speed and constant static
pressure, resulting in a constant flow. As a result, the
fan can be operated at a condition of maximum efficiency.
The velocity of the mainstream airflow is controlled by the
balance between the mainstream outlet vane and the bypass
outlet vane.
The bypass ducting splitter must be carefully
designed because flow disturbances and distortions of the
mainstream velocity profile can occur. This is particularly
true if a single bypass duct is used as shown in Figure 5-15a
In this configuration the bypass air is required to make a
90° turn, and, as a result, the velocity profile of the
mainstream air supply will be skewed as shown in Figure 5-15.
Some smoothing of the flow profile can be achieved by using
flow straighteners or a noise baffle, but the inclusion of
these devices increases the pressure drop and the size of
the system. Also, if a closed-loop control system with a
velocity sensor is used, the system response time would be
increased because of the added length of the ducting.
One configuration which minimizes the flow distur-
bances is shown in Figure 5-15. In this design, two bypass
ducts are employed, and the flow is symmetrically divided
about the center line of the ducting. Some distortions in
the flow will still occur, and the resulting profile will be
an exaggerated parabola. This effect can be reduced by
utilizing an opposed blade damper whose multiple vanes will
function as a flow mixer. Use of this style of damper would
also result in a nearly linear control mechanism. Since the
5-40
-------
a. SINGLE-DUCT BYPASS
•HINGES
b. DUAL-DUCT BYPASS
Figure 5-15. BYPASS VANE CONTROL SYSTEM
5-41
-------
blower would always be operating at a constant condition,
the dampers can be equipped with special vane-edge wiper
seals in order to achieve the dynamic airflow range required
by the specification.
Depending on the precision and linearity of dampers
and related linkage system, this approach can be easily used
in either a closed-loop or open-loop system. The cost of
the system is estimated to be $4,000 ($1,000 for dampers and
ducting, $1,500 for controls and linkages, and $1,500 for
feedback sensor or precision controller).
There are several disadvantages to this design
approach. First, the blower is always operating at its
maximum capacity, which is three times the average required.
According to the fan laws, horsepower varies as the cube of
air volume, as shown in equation (5-2):
new BHP = old BHP x
/new CFMV
1 old CFM I
Therefore, since the maximum LA-4 test speed is three times
the average test speed a fan operating in a bypass configura-
tion will consume approximately 27 times more power than a
damper controlled or variable-speed controller system during
a LA-4 cycle. In addition, in this design, two control
mechanisms would have to be operated simultaneously, although
they can be interconnected. Average sound emissions would
also be high with this system since the blower is always
operating at a peak condition. Finally, no suitable commer-
cially available control vanes or valves have been located
for this application, and the special design of suitable
mechanisms could result in added design efforts and project
costs.
5-42
-------
5.6.2.3 Variable-Speed Drive Control
Variable-speed drive control is the most expensive
means of achieving a variable-flow blower system. The costs
for this system excluding a feedback sensor would be $5,000
to $8,500 depending on the size of the motor, the electronic
braking system, and the system's stability and response
time. However, this approach is proven and would be capable
of satisfying the requirements outlined in Table 5-1.
According to the fan laws, which approximately define the
characteristics of a fan operating under differing conditions,
new CFM = old CFM x ' new RPM
[ ne
V"
d RPM / (5-3)
and
new BMP - old BMP x R™
R . (5.4)
Since flow varies proportionately with rotational speed of
the motor or fan wheel, a variable-speed motor control
system is approximately a linear system which can be utilized
in either open-ended or closed-loop control systems.
Furthermore, as the rotational speed is being reduced, the
required horsepower decreases cubically resulting in a
substantial reduction in consumed power and component degrada
tion. By controlling the RPM of the blower rather than its
static pressure, operation of a blower under conditions of
instability or stall is also avoided.
In a motor speed control system the motor selected
must be able to follow the acceleration/deceleration rates
outlined in the driving cycle in addition to the requirement
of rotating the fan wheel at the maximum rotational speed.
5-43
-------
The power required to achieve an angular acceleration rate
is defined by the expression:
P = I Ar Vr (5-5)
where
I = rotational inertia of the fan wheel
A = rotational acceleration of the fan v»
V = rotational velocity of the fan wheel
Since the rotational inertia of a vaneaxial blower
is almost an order of magnitude less than that of a centrifu-
gal blower's impeller, less power will be required to acceler-
ate the former even though its angular acceleration rate is
higher. It has been estimated that for a suitable vaneaxial
blower, an additional 4 hp would be required to meet the
acceleration requirements imposed by the driving cycle. An
additional 10 hp would be required for the operation of a
centrifugal blower.
Utilization of a variable-speed motor to achieve
airflow control during the mileage accumulation tests will
require a motor almost 50 percent larger. The added power
will be required to achieve the higher maximum speed and
match the wide open throttle acceleration which occurs
during lap 11 of the mileage accumulation test.
Two different types of variable speed drives are
commercially available and adaptable for this application.
These are the constant-speed AC motor with an eddy current
clutch and a DC motor system.
Two manufacturers, Louis Allis and Eaton Dynamatic,
have supplied literature describing similar eddy current
systems and a typical drive is shown in Figure 5-16. In
this drive a standard three-phase induction motor drives the
input member of the eddy current clutch at a constant speed.
5-44
-------
Stationary field
clutch coil
Annular grooved drum for
increased surface area
Over temperature
switch
Terminal board
connections
\ Blower with
HU-Louis Allis
TEFC motor
Tachometer generator
with encapsulated coil
Non- magnetic
flux barrier
Pressurized cooling
chambers
Figure 5-16. EDDY CURRENT DRIVE
5-45
-------
The stationary field coils in the clutch generate magnetic
fields which, in turn, generate eddy currents in the input
member drum. These eddy currents generate magentic fields
which expose the primary fields and give rise to an electro-
motive force between the AC motor and the load. Speed
control is maintained by modulating the DC excitation current
to the clutch coil. A basic drive control loop is shown in
Figure 5-17. The electronic control must include current
feedback curcuits for stability, current limiting circuits
for coil protection, regulated power supplied, and isolation
transformers for safety, reliability and noise immunity.
A 20 hp drive would be required to operate a
vaneaxial blower required for this application, and this
drive could deliver sufficient torque to meeet the accelera-
tion demand of the LA-4 cycle, but not necessarily the WOT
conditions of the eleventh lap of the mileage accumulation
cycle. During the deceleration modes of the LA-4 cycle,
supplementary braking would be required. With this type of
drive, 0.5 percent requlation can easily be maintained over
a 35-to-l speed range. The approximate price of an Eaton
unit, Model 324T is $6,500 with supplementary braking. The
Louis Allis equipment is comparably priced.
There are several drawbacks to this system. First
the dynamic range of this system is only 35-to-l as compared
to the specified range of 100-to-l. However, this is not a
serious deficiency since it is doubtful that a motorcycle
being cooled by a 3 km/hr breeze would exhibit significantly
different emissions than one cooled by a 1 km/hr wind. A
more serious drawback is the response time of the system
estimated to be between 0.5 and 1.5 seconds and the power
needs of the drive. Since the induction motor is always
operating at a constant speed, it will usually be drawing
more power than is required to operate the blower. This
unused power will be dissipated into the room in the form of
heat, increasing the load on the room's heating and air
conditioning system.
5-46
-------
SPEED
REFERENCE
r
(ERROR DETECTOR)
4-
SUMMI
<\MPLI-
FIE
I
ELECTRONIC CONTROL
VELOCITY FEEDBACK
( MOTOR J
r
1C
COIL
EDDY
CURRENT
| CLUTCH
MECHANICAL
LOAD
TACHOMETER
GENERATOR
Figure 5-17. BASIC VELOCITY CONTROL LOOP
EDDY CURRENT DRIVE
5-47
-------
In spite of these drawbacks, the eddy current unit
is well suited for this application. It is a proven, turn-
key system which has been utilized in similar applications.
While its cost is greater than the bypass approach, it has
already proven itself and can be quickly and easily adapted
for this application with almost guaranteed satisfaction.
Two types of variable-speed DC drives are available,
The first is equipped with a SCR/thyristor controller and
the other style employs a regenerative type controller. The
latter unit has a wider range of operation, including braking
and greater regulation than the former, but there is a cost
penalty. Compartive performance data is summarized in
Table 5-5.
A representative SCR system, the Max Pak V*S Drive
is produced by Reliance Electric Company, and consists of
three major components.
• A variable-speed DC motor designed for opera-
tion from a phase controller rectified power
supply.
• A solid-state power converter and regulator
for providing rectified power.
0 An operator's control station, which contains
the master speed control and the pushbutton
control devices to operate the driver.
A typical interconnection diagram is shown in Figure 5-18.
The control unit performs three separate tasks.
Its primary function is to rectify the input AC current and
regulate the DC output. This is achieved using SCR/thyristor
circuitry. The second function of the controller is to
provide protection from AC surges, instabilities, etc. The
final element is the operator's control panel which contains
5-48
-------
0« H ATOM
SPIED
CONTROL'
Figure 5-18. TYP|CAL INTERCONNECTION DIAGRAM
5-49
-------
the master speed control potentiometer, jog-run selector
switch, and start-stop pushbuttons.
The variable-speed DC motor has been designed to
operate from rectified power and can be modified to include
gearmotors, tachometers, slide bases, etc. While motors can
be supplied in a variety of base speeds, the most common
speed is 1,750 rpm. A substantial cost premium, approximately
$1,000 must be paid if a slower speed motor is required.
Typically, the drive can be operated at speeds as low as
l/20th of the base speed without providing supplementary
cooling for the motor. However, this cooling is mandatory
when the motor speed is less than 5 percent of the base
speed for sustained periods.
Typically, DC drive systems can provide regulated
speed control within ±2 percent over an operating range of
50-to-l. However, optional circuitry can be acquired to
improve the regulation to ±0.1 percent and increase the
range to 100-to-l. As in the eddy current system, the DC
drive can easily meet the acceleration demands imposed by
the LA-4 cycle, but no braking is provided to meet the
deceleration demands.
The regenerative system is basically the same as
the previously described unit. However, additional controlled
rectifiers have been added to accomplish regeneration and
reversible operation. As a result of the added circuitry,
the regenerative system has a rapid response time, less than
500 ms, a wide dynamic range of operation, and does not
require supplementary braking. A flow control system which
utilizes a Randtronic's regenerative unit is schematically
shown in Figure 5-19. This system is somewhat different in
that it employs a dual feedback arrangement, using both a
tachometer mounted on the motor shaft, and a air velocity
sensor. The price of this system is $8,300 and delivery
would be 14 weeks after the receipt of an order.
5-50
-------
0-15 VDC
0-5 VDC
ROAD
VELOCITY
DYNO
MOTOR
480 V
3 0
60 Hz
en
i
en
ANEMOMETER
0-15 VDC
TB 700 -
140-240
CONTROLLER
-240 -0 +240 VDC
40 HP
1750 RPM
0-5 V
MAX ACCEL RATE
TO DYNO DRIVE COMMAND
= 5.3 Km/Hr/Sec
max
Km/Hr
-WIND VELOCITY
Figure 5-19. DC REGENERATIVE CONTROL SYSTEM
-------
5.6.2.4 Compound Control System
A compound control system would employ two or more
of the control methods described above. Therefore, it would
be more complex than any system employing a single control
mechanism. As a result of its more complex nature, the
development of a compound system should not be pursued
unless no single control mechanism is capable of meeting
requirements outlined in Table 5-1.
Since a single control method appears to meet the
design objectives of this application, further examination
of the more complex compound system is unwarranted at this
time.
5.7 RECOMMENDATIONS
5.7.1 Ranking of Alternatives - Blowers
Optimum Selection
1. Centrifugal - Airfoil Impeller
2. Centrifugal Blower - Backward Inclined Impeller
Acceptable Selection
3. Vaneaxial Fans
Unacceptable Selection
4. Mixed Axial Centrifugal Blower
5. Centrifugal Blower - Radial Impeller
Not Applicable
6. Tubeaxial Fans
7. Centrifugal Blower - Forward Curved Impeller
8. Propeller Fans
5-52
-------
As a result of the investigation conducted to
date, no discernible differences between centrifugal blowers
equipped with airfoil blades and backward curved blades
could be identified, and it is felt that these units are
directly interchangeable. Further consideration should be
given to a vaneaxial unit because of its more compact design,
low weight, low rotational inertia, uniform outlet air
distribution and low cost.
The mixed axial unit and the centrifugal blower
equipped with radial blades were considered marginally
acceptable because of their high cost, large size, poor
efficiency and high noise emission. The other blowers were
disregarded because of their inability to satisfy the require-
ments outlined in Table 5-1.
5.7.2 Ranking of Alternatives Variable Flow Control Methods
Acceptable
1. Eddy Current Clutch with AC Induction Motor
2. Bypass Control - Dual Bypass Ducts
3. DC Regenerative Drive
4. SCR DC Drive
Marginally Acceptable
5. Bypass Control - Single Bypass Duct
Unacceptable
6. Inlet Dampers
7. Outlet Dampers
Of the four acceptable choices, the eddy current
drive is recommended. While its cost is somewhat greater
than the bypass control alternative, it is an off-the-shelf
system which is commonly used in similar applications.
While the eddy current drive does not have the range and
5-53
-------
degree of regulation of the two DC drives, it is felt that
its performance is sufficient to satisfy the system objectives,
although it may not fully meet the design specifications set
out in the discussion. The regenerative drive is recommended
over the standard SCR system because of its increased regula-
tion, range, response time and braking capabilities.
The dual duct bypass control is definitely superior
to the single duct unit, even though it is somewhat costlier
and more complex.
While shut-off dampers provide the least expensive
alternative, their dynamic range is insufficient to fulfill
the requirements of the project.
5.7.3 System Recommendations
It is recommended that two design alternatives be
developed further in Task 6 and that a specification based
on the most promising of these designs be written.
The first design would employ a vaneaxial fan and
a variable speed drive, probably an eddy current unit, but
possibly a regenerative DC drive. The vaneaxial blower was
selected for use with this drive because its inherent
performance/design characteristics; i.e., low rotational
inertia mass; maximum rotation speed of 1,750 RPM; output
proportional with rotational speed; match the needs of the
drive. The increased costs of the variable-speed drive are
offset by the lower costs of the vaneaxial blower. Noise
emissions from this system should not be offensive, since
the blower will only be operating at maximum flow conditions
for a short period. Supplementary cooling of the drive will
not be required because the motor will be directly coupled
to the fan wheel and will be continually cooled by the inlet
air passing over it.
The second system is a centrifugal blower mounted
in dual-duct bypass configuration. The centrifugal blower
5-54
-------
is well suited for this mode of operation because of its
high efficiency, low noise output, ability to overcome
substantial (5 to 8" H20) pressure drop, and high rotational
inertia which provides stable operation over prolonged
periods of time.
5-55
-------
Section 6
TASK 6 - COOLING SYSTEM ANALYSIS
6.1 INTRODUCTION
The design of a variable-speed blower system which
simulates road-cooling conditions will be detailed in this
report. The requirements as set forth in the EPA Draft
Regulations and the contract will be reviewed, and the
recommendations of the Task 5 Report, as modified by the EPA
Project Officer, will be summarized. Additionally, in this
report a cost-trade-off analysis of the design alternatives
will be conducted.
6.2 REVIEW OF REQUIREMENTS
Paragraph 85.478-15 (b) of EPA's Notice of Proposed
Rulemaking (NPRM) for New Motorcycles, specifies the incor-
poration of a variable-speed blower system for simulating
engine cooling. The Regulations and the contract require
that the blower have a mechanism, controlled by the dyna-
mometer roll speed, which will regulate the blower flow to
within 10 percent of the roll speed over, an operating range
of 10 kph (6.2 mph) to 100 kph (62 mph). At speeds less
than 10 kph the air flow shall be within ±1 kph.
The NPRM and the contract also describe the config-
uration and positioning of the blower outlet duct. As
stated in the contract, "the cooling system shall be located
to discharge air to the front of the motorcycle so that the
movement of air over the motorcycle engine simulates the
6-1
-------
movement of the motorcycle through the air. The cooling
system shall have a square or rectangular outlet of at least
0.5m2 (5.38 feet2) outlet area. The lower edge of the
outlet shall be located about 150mm (5.91 inches) above
dynamometer floor level. Outlet flow shall be uniform to
±20 percent across the outlet area as measured at the
center of the area as compared to the center of each quarter
area."
In the presentation which follows, the rationale
behind these requirements will be reviewed, and the feasi-
bility and/or desirability of achieving these objectives
will be -examined.
6.3 REVIEW OF TASK 5 REPORT RECOMMENDATIONS
In the Task 5 Report, separate rankings for blower
types and flow control mechanisms were presented. The
acceptable alternatives are listed in Table 6-1. Also in
that task report two alternative systems were recommended
for further development. The first design utilized a vane-
axial fan with a variable-speed drive such as an eddy current,
DC, or regenerative DC motor. The second system employed a
centrifugal blower mounted in a dual-duct bypass configura-
tion.
These findings were discussed in detail at a
review meeting held with the Project Officer and other
members of the EPA staff on May 21, 1975. On May 29, 1975
the Project Officer directed Olson Laboratories to examine
only the following two alternatives in the Task 6 Report:
1. A vaneaxial blower used in conjunction with
a variable-speed drive.
2. A centrifugal blower used in conjunction with
a variable-speed drive.
6-2
-------
Table 6-1. TASK REPORT 5 - RECOMMENDATIONS
A. Blowers
Optimum Selection
1. Centrifugal Blower - Airfoil Impeller
2. Centrifugal Blower - Backward Inclined
Impeller
Acceptable Selection
3. Vaneaxial Fans
B. Flow Control Methods
Acceptable
1. Eddy Current Clutch with A-C Induction Motor
2. Bypass Control - Dual Bypass Ducts
3. DC Regenerative Drive
4. SCR DC Drive
Marginally Acceptable
1. Bypass Control - Single Bypass Duct
6-3
-------
Olson Laboratories concurs that these two systems
are viable alternatives, which would result, with a minimum
design effort, in systems which achieve the design goals
outlined in Section 6.2.
6.4 DESIGN ALTERNATIVES
As stated in Section 6.3, the Project Officer has
requested that two design alternatives be examined in detail.
As a result of this study, a final design will be developed
and specifications will be formulated. In this section of
the task report, evaluation criteria will be defined and a
cost-effectiveness study of the design alternatives will be
conducted. The alternatives will be ranked and a recom-
mendation made.
In reality there are six system designs which
should be considered. These combine the two types of blowers
with the three types of variable-speed drives. The six
combinations are:
1. Vaneaxial blower with eddy current drive.
2. Vaneaxial blower with DC drive.
3. Vaneaxial with regenerative DC drive.
4. Centrifugal blower with eddy current drive.
5. Centrifugal blower with DC drive.
6. Centrifugal blower with regenerative DC drive.
Additionally, there are several suppliers for each
component. Characteristics of representative units have
been utilized in this analysis. However, specific design
features which either greatly improve or degrade a compon-
ent's performance in this application will be noted, and the
specification will appropriately be revised. In no case
will the specification be so limited that only one supplier
can meet the requirements.
6-4
-------
6.4.1 Ranking System
In order to evaluate the alternative designs, a
number of evaluation criteria have been established. These
have been divided into two sections; the first discusses
criteria related to the blower, and the second addresses the
motor control system and the resulting system performance.
All of the alternatives will be evaluated with
respect to each of the criteria and will be rated on a scale
of 1 to 5, where 1 is unacceptable, 3 is acceptable, and
5 is outstanding. Additionally, a weighting factor has been
designated for each of the criteria, defining its relative
importance. Each of the ratings will be adjusted by using
the weighting factor, and the rankings of the alternatives
will be calculated by summing the weighted ratings.
The evaluation criteria with their respective
weighting factors are listed in Table 6-2, as well as the
unweighted ratings.
6.5 EVALUATION CRITERIA - BLOWER
6.5.1 Flow as a Function of RPM and AP
In order to define the blower control system, the
relationship between the blower output and the rotational
speed of the impeller must be established. Ideally, this
relationship should be linear; thus permitting a simple
tachometer control circuit. If the relationship is non-
linear, then, as reported in the Task 5 Report, a velocity
sensor must be used as the feedback sensor in the control
system which will result in a more complex control circuit
and ducting arrangement.
The fan laws state that
new CFM = old CFM x newjrpm (6-1)
old rpm
6-5
-------
Table 6-2. UNWEIGHTED RATINGS
CRITERION
I. Blower Performance
1. Flow as a function of rpm
and AP
2. Rotational Inertia
3. Delivery
4. Noise
5. Reliability/Maintenance
6. Adaptability
7. Power Consumption/
Efficiency
8. Motor Compatibility
9. Size and Weight
10. Flow Conditioning
11. Cost
II. Motor Control and
System Performance
1. Accuracy
2. Range
3. Response Time
4. Delivery
5. Size and Weight
6. Adaptability
7. Reliability/Maintenance
8. Interferences
9. Cost
WT.
FACTOR
1.0
0.9
0.8
0.8
0.8
0.8
0.7
0.7
0.6
0.5
0.5
1.0
0.9
0.9
0.8
0.7
0.7
0.7
0.7
0.7
SYSTEM NO. (SECT. 2.2)
VANEAXIAL BLOWER
EC DC RDC
555
333
333
333
444
333
233
322
444
444
444
4 1 4
244
3 1 3
333
333
4 1 4
433
333
333
CENTRIFUGAL BLOWER
EC DC RDC
222
222
333
444
333
333
244
444
333
222
333
2 1 2
1 1 1
2 1 2
333
333
2 1 2
433
333
333
CTl
I
-------
and
/ \0
(6-2)
new static pressure (SP) = old SP xT^ ^
\ t
However, these formulas are most accurate when
used to define conditions at two operating points which are
fairly close together. Under the conditions imposed by this
application, the blower will be operating over a 100-to-l
range, and over this range, the validity of the fan laws is
marginal .
In order to establish flow versus rpm curves for
vaneaxial and centrifugal blowers of the appropriate capacity,
the static pressure requirements for this application, as a
function of flow, must be defined. The static pressure
which the blower must overcome consists of two components.
The first occurs as frictional losses in the ducting and is
a function of the ducting configuration, ducting smoothness,
and mean air velocity. The second component results from
pressure losses occurring as a result of contractions and
expansions in the duct cross-section. The magnitude of this
loss is affected by the ratio of the two duct areas and the
volumetric flow of air.
A family of curves has been plotted in Figure 6-1
defining the total static pressure loss as a function of
blower exit area for several volumetric flow rates. For the
purposes of those calculations, the duct exit area has been
assumed to be 0.5m2 (5.38 feet2).
Utilizing performance data supplied by manufac-
turers and the static pressure demands defined in Figure 6-1,
performance curves relating air delivery and rotational
speed were developed for four blowers, two vaneaxial units
and two centrifugal units. These curves are shown in Fig-
ure 6-2. Performance data, supplied by the manufacturer,
describe the operation of a blower in a typical industrial
application, in which the blower operates at a specified
6-7
-------
condition or over a confined range of conditions, usually no
more than a 10-to-l range. In this application, the blower
will be required to operate aver a 100-to-l range. There-
fore, in order to define operation at slower speeds, the
manufacturer's data has been extrapolated.
In the case of the centrifugal units, this extra-
polation has provided some valuable information. Referring
to Figure 6-2, the larger centrifugal unit displayed a
linear relationship between rotational speed and air flow
over a 3-to-l range of operation, while the smaller unit
appeared to be linear over a 2-1/2-to-l range. When these
performance curves were extrapolated down to a no-flow
condition, however, a negative rotational speed resulted.
This, of course, does not occur in the real world. There-
fore, since the performance curve must go through the origin
of this plot, it is apparent that the performance curve of
the centrifugal blower is not linear over the entire range
of interest. In contrast, the vaneaxial fans performance
curves are linear and do appear to go through the origin
when the performance curve is extrapolated to a zero-flow
condition.
Based on these curves, the vaneaxial blower
appears to have a distinct advantage over the centrifugal
unit when either an open-loop or closed-loop tachometer
control system is used.
6.5.2 Fan Wheel Rotational Inertia
The inertia of the blower's driven element at the
motor shaft is an important parameter affecting the selec-
tion of the motor and control system. The motor must exert
an accelerating torque sufficient to overcome the fan wheel's
inertia and accelerate it to the operating speed. The
acceleration rate is a function of the magnitude of the
inertia.
6-8
-------
600--
X
'(30,000 CFM)
500 --
400 - -
(0
O.
LU
CC
n
co
CO
LU 300
or
o
co
200--
100--
0
0
(25,000 CFM)
(20,000 CFM)
(15,000 CFM)
(10,000 CFM)
0.5 1.0 1.5
FAN OUTLET AREA (m2)
2.0
FIGURE 6-1. STATIC PRESSURE VS. FAN OUTLET AREA
6-9
-------
1.0- -
CTl
I—1
O
§ 5/6
u.
O
o
o
o
2 2/3
o
Ui
CO
2
£ 1/2
•*'
II
x
2
O
1/3- -
1/6- -
/ / ,',' ''
/ / ,'S ,' BLOWER
/ / ''' ''
/ / ^s''' ,'' BARRY 490
,','' ,'' BARRY 402
,/s' ,'' BUFFALO 38A5
,'^' ,'' BUFFALO 38A9
,'2''''' BUFFALO 43B5
's's' 1 1 1 1 1
TYPE
CENTR
CENTR
V.A.
V.A.
V.A.
1
INERTIA
(kgm2)
30.4
14. 9
3.4
3.6
6.2
1
OUTPUT
AREA
1.28
0.86
0.74
0.74
0.97
— 1 1
200
400
600 800 1000
FAN WHEEL R.P.M.
1200
1400
1600
1800
FIGURE 6-2. FAN FLOW VS. ROTATIONAL SPEED
-------
Data on acceleration rates as a function of fan
wheel inertia were obtained for both eddy current and DC
motors. This data is summarized in Table 6-3.
Table 6-3. FAN WHEEL ACCELERATION RATES
(Zero to Max RPM)
BLOWER
Hartzell
VA 36A
Barry
445 SISW
New York
Bl ower
449 SISW
MAX
RPM
1,750
800
890
INERTIA
(kg m2)
2,73
21.30
25.40
EC MOTOB
ACCEL1
(sec)
1.9
3.1
3.6
DC MOTOR
ACCEL^
(sec)
2
5
8
Data supplied by Louis Allis Co. 40 hp Adjusto-
Speed Motor.
>
'Data supplied by Robicon Corp. 40 hp General
Electric Motor.
As shown in the table, the vaneaxial Hartzell blower can be
accelerated to its maximum speed faster than the centrifugal
units even though its maximum rotational speed is twice that
of the centrifugal units. This demonstrates the importance
of fan wheel inertia.
As shown in Table 6-3, there appears to be a
difference in the acceleration characteristics of the two
types of motor controls. However, motor vendors indicate
that acceleration rates are dependent on the motor design,
specifically the armature inertia. Therefore, for each
class of motors, various acceleration rates are available.
Achievable acceleration rates are greater than those required
to meet the objectives of this program. In the specification,
a minimum acceleration rate and response time are imposed by
the accuracy specification.
Deceleration rates will also be dependent on the
inertia of the fan wheel. However, deceleration times with-
out the use of external braking or regenerative techniques
6-11
-------
will be long, probably in excess of a minute, and unaccept-
able for this application. Braking requirements will be
discussed in Section 6.6.3.
6.5.3
Delivery Schedule
Time required for delivery of the system is con-
sidered important because it could impact on the EPA's plans
to institute a certification program on 1978 motorcycles.
Several manufacturers have been contacted to determine
representative delivery schedules. These are summarized in
Table 6-4. As shown in this table delivery dates are
reasonable, and there appears to be no distinct differences
between deliveries of vaneaxial and centrifugal units.
Table 6-4. TYPICAL DELIVERY INFORMATION
MANUFACTURER
Barry Blower
New York
Bl ower
Aerovent
Buffalo
Forge
MODEL
NO.
402
455
490
449
449
V361
V381
33A5
33A9
38A5
ARRANGE-
MENT NO.
7 SWSI
7 SWSI
7 SWSI
8 SWSI
8 SWSI
Type W
Type W
4 Type S
4 Type S
4 Type S
DELIVERY
(days)ARO
90
90
90
90
90
70
70
125
125
125
$
1509
1830
2171
2360
2964
1041
1219
875
1058
956
SHIPPING
WT. (kg)
544
680
771
742
952
272
327
142
142
166
6.5.4
Noise
Noise levels within the test cell must be kept at
tolerable levels. As discussed in the Task 5 Report, expected
noise levels from a vaneaxial unit will be greater than
those for a comparable centrifugal unit. A comparison of
the expected noise levels is shown in Table 6-5. Sound
6-12
-------
Table 6-5. NOISE LEVELS
BLOWER TYPE
New York Centr.
Blower 449
New York Centr.
Blower 490
Aerovent V.A.
V361-Y34
Aerovent V.A.
V381-Y30
New York Centr.
Blower 449
New York Centr.
Blower 490
Aerovent V.A.
V361-Y34
Aerovent V.A.
V381-Y30
20 - 75 -
75 Hz 150 Hz
98 101
103 96
101 94
102 94
83 86
88 81
86 79
87 79
SOUND
150 -
300 Hz
9 835
91
91
95
95
@ 417
76
76
80
80
POWER LEVELS dB re
300
600
3, .
m /mm.
89
85
98
99
m /min
74
70
83
84
600 -
Hz 1200 Hz
(29,500 cfm)
87
84
98
98
. (14,750 cfm)
72
69
83
83
-12
10
1200 -
2400 Hz
80
81
96
95
65
66
81
80
WATT
2400 -
4800 Hz
74
76
88
89
59
61
73
74
4800 -
10000 Hz
70
74
82
82
55
59
67
67
dBA1
90
84
88
87
75
69
73
72
*dBA estimated at 1.5 m from blower outlet
-------
power levels and dBA levels have been reported in the table
for two different flow conditions, maximum flow and one-half
maximum flow.
The dBA noise levels reported in Table 6-5 do not
exceed OSHA standards which permit an individual to be
exposed to sound levels 90 dBA and below for periods longer
than 8 hours. Some care must be taken by the facility
designer, however, to protect test-cell areas from the high
frequency noise emitted by the vaneaxial unit.
The estimated sound magnitudes do not warrant the
installation of sound attenuators. While attenuators would
reduce noise levels significantly, they introduce added
static pressure losses which would increase horsepower
requirements and costs. Maximum sound levels of 90 dBA and
average sound levels of 75 dBA are tolerable, and well
within acceptable limits.
6.5.5 Reliabili ty/Maintenance
The reliability and maintenance requirements for
the two blower types were examined. Both units are highly
reliable because of their inherent simplicity.
The vaneaxial blower, excluding the motor, requires
no maintenance since the motor is connected directly to the
impeller wheel and the blower itself has no bearings. This
is in contrast to the centrifugal blower which utilizes two
bearings, one on each side of the impeller to support the
impeller wheel. These bearings are typically grease-
lubricated, heavy-duty and self-aligning. The average
service life of these bearings is 125,000 hours using B-10
minimum-1ife-rating data as defined by the Anti-Friction
Bearing Manufacturers Association. The minimal servicing
required for the centrifugal blowers should not preclude
their use in this application.
6-14
-------
6.5.6 AdaptabilIty
The system which is being specified in this report
has been designed around the requirements imposed by the
Federal Test Procedure (FTP), but it also should be flexible
enough to satisfy other potential requirements. These added
requirements could require increased flow capacity which can
be achieved easily with either style of blower. However,
all techniques for additional flow require more power.
Added flow output from a centrifugal blower can be
achieved by increasing the rotational speed of the impeller.
This is easily accomplished because motors which are readily
available for this application have a maximum operating
speed of 1,170 or 1,750 rpm, which is substantially faster
than the 600 to 800 rpm needed to satisfy the basic require-
ments imposed by the FTP. Before selecting a centrifugal
blower, the compatabi1ity of the blower's structural design
and the maximum rotational speed required should be verified,
since all blowers have maximum operating speeds. As an
example, Table 6-6 defines the maximum safe rpm of the New
York Blower units at ambient conditions.
Table 6-6. MAXIMUM SAFE WHEEL RPM
OF SWSI BLOWERS
SIZE
409
449
449
CLASS I
1005
910
850
CLASS II
1315
1190
1105
CLASS I
1655
1495
1395
II
Ref: Bulletin 725 B New York Blower
Company
The output from a vaneaxial blower can also be
increased by operating the Impeller wheel at faster rota-
tional speeds. Typically, however, these units are usually
operated and rated at a rotational speed of 1,750 rpm which
6-15
-------
is the rated maximum speed of standard variable-speed motors
Increased flow from vaneaxial units can also be achieved by
changing the pitch of the blower vanes. This effect can be
seen in Figure 6-3 which describes the performance of a
vaneaxial blower manufactured by Buffalo Forge. In this
style blower, operating at 1,750 rpm, a 280 percent increase
in flow can be achieved by adjusting the pitch of the blades
Almost all manufacturers of vaneaxial units have adjustable
pitch wheels which permit the user to set manually the pitch
of the blades. Two manufacturers supply units in which the
pitch can be adjusted pneumatically or electrically while
the fan is in motion. However, these units do not have
sufficient dynamic range to meet the 100-to-l variations
required for this application.
It is concluded that both blowers can be easily
adjusted for increased flow operation if the motors and
blower are carefully selected.
6.5.7 Power Consumption/Efficiency
The mechanical efficiency (ME) of a fan is defined
by equation (6-3).
ME = —emir <6-3)
where Q = volumetric flow from blower (m3/min)
Py = total pressure (Pa)
W = shaft power delivered by motor (KW)
The average mechanical efficiency of the blower system
during one Federal Test Cycle is:
6-16
-------
BIADE SETTING
10 i>0 [ID '10 -M (iO
- 20
\lj TYPE S, ADJUST AX "•
" Vr. 9 VANEAXIAL FAN
SIZE 33A9
20 30
CAPACITY - C.F.M. x 1000
40
50 60
FIGURE 6-3. VANEAXIAL BLOWER PERFORMANCE
6-17
80
-10
- 8.0
- 6.0
- 5.0 „
- 4.0
- 3.0
2.0 <
o
- i.O
-O.c
- C.6
-------
ME
1371
Q(t) PT (t)
672 W(t)
-dt (6-4)
As shown in equation (6-4), the volumetric flow,
total pressure and shaft power are all functions of time;
varying as the motorcycle speed varies. Also the total
pressure and shaft power functions will be different for
each b'lower under consideration. Establishing the effi-
ciency of the total system is further complicated because
the efficiency of the drive motor should also be considered,
and that varies with shaft rpm. Also, the function W(t)
cannot be defined due to lack of suitable data. Blower
manufacturers have not developed the necessary data because
operation of a blower at 1 percent of its maximum speed, is
a condition which is never encountered by the average user.
Therefore, rather than explicitly calculating
equation (6-4) for each blower configuration and size, a
semi-quantitative study has been conducted, comparing the
power requirements and efficiency of several blowers at
different air volume outputs. The results have been plotted
in Figures 6-4 and 6-5.
In Figure 6-4, the power requirements of several
blowers are examined. On the average, the vaneaxial units
require more power than the centrifugal units, and the power
required by the vaneaxial unit appears to be independent of
the blower size. In contrast, the power required to operate
a centrifugal blower appears to decrease as the blower size
increases.
The efficiency of the various blowers are compared
in Figure 6-5. As shown in the plots, vaneaxial units are
6-18
-------
14 --
12 --
10 •-
* 8
QC
LLJ
$
o
Q.
DC
O 6
H
o
4 ._
2--
0
BARRY 402 SWSt
VANEAXIAL BLOWER
200 400 600
BLOWER OUTPUT (M3 /MIN.)
800
YORK BLOWER
449 SWSI
NEW YORK BLOWER
499 SWSI
1000
FIGURE 6-4. BLOWER POWER REQUIREMENTS
VS. BLOWER TYPE AND SIZE
6-19
-------
100 •-
80'-
-------
less efficient than comparable centrifugal units.
When considering system efficiency and power
consumption, one must evaluate the motor as well as the
blower. Efficiency estimates have been obtained for both
eddy current and DC motors. The DC motor is not as effi-
cient at maximum speeds as the eddy current unit. The
former is 70 percent efficient while the eddy current unit
exhibits an efficiency over 90 percent. However, as the
rotational speed decreases the situation reverses. The
efficiency of the DC motor remains constant over the motor's
speed range, but that of an eddy current unit drops to about
40 percent at half speed. At quarter speed, its efficiency
is less than 25 percent.
6.5.8 Motor Compatibility
As proposed in the Task 5 Report and reconfirmed
in this report, a direct-drive motor should be used to power
the blower. This requires that the blower and motor be
physically compatible. The use of a direct-drive motor with
each of the blower styles under consideration imposes
certain design considerations and limitations. These will
be subsequently discussed.
Centrifugal blowers are available in several
direct-drive configurations. These are shown in Figure 6-6.
Arrangements 1, 2, and 3 require the user to fabricate a
motor base and interface the motor and the blower. Arrange-
ments 4, 7, and 8 provide a motor base designed by the
blower manufacturer to mate with the specific drive being
used. The difference between these latter three arrange-
ments is the use and positioning of bearings. Arrangement 4
utilizes no bearings, while two bearings are used in the
other two arrangements. For this application, Arrangement 7
is superior. Its use of two bearings, one on each side of
the fan wheel, provides excellent support for the unit and
maximizes bearing life and utility.
6-21
-------
ARRANGEMENT 1
Two bearings on base
Wheel overhung
No motor base
ARRANGEMENT 2
Bearings in bracket
supported by housing
Whee 1 ove rhung
No'motor base
ARRANGEMENT 3
One bearing on each
s ide of whee1
No motor base
ARRANGEMENT k
No bearings on fan
Wheel overhung on
motor
Motor base
ARRANGEMENT 7
Same as Arrangement
plus motor base
ARRANGEMENT 8
Same as Arrangement
plus motor base
FIGURE 6-6. CENTRIFUGAL BLOWER MOTOR
MOUNTING CONFIGURATIONS
6-22
-------
With the centrifugal blower configurations shown
in Figure 6-6, there is no limitation to the size of the
motor used. Either drip-proof or totally enclosed motors
can be used. The only precaution necessary is that the
blower and motor base be securely fastened to the test cell
f1oor.
Motor size is a critical parameter in the design
of a vaneaxial fan system. In a direct-drive vaneaxial fan,
the motor is close-coupled to the fan wheel and sits within
the fan housing. The motor diameter must be smaller than
the hub diameter of the fan wheel. Otherwise, the motor
will interfere with the movement of air within the fan; thus
decreasing its efficiency and capacity. Table 6-7 details
the maximum motor frame size which can be accepted by a
representative group of vaneaxial fans, and it also defines
the frame size of the variable-speed motors being considered.
As shown in Table 6-7, the available motors barely fit
within the fan housings. If a totally enclosed motor is
required, the fit is even tighter and the selection of
compatible motors and fans is severely limited.
The compatibility of fan and motor is further
affected by the specific design of the fan. As an example,
the Buffalo fans have been designed for use only with motors
having C-faced flanges (no feet). Although this option is
available from the DC motor manufacturers, it is not compat-
ible with eddy current motors where feet are required to
reduce the stress at the motor face.
Motor design is another parameter which must be
considered. Buffalo Forge only recommends motors that are
totally enclosed with air-over construction. This design
prevents the free interchange of air between the inside and
outside of the motor, resulting in excellent protection to
all internal motor components. Hartzell Propeller also
recommends the use of totally enclosed motors. However,
Aerovent states open-type motors can be used, if they have a
6-23
-------
Table 6-7. MOTOR/VANEAXIAL FAN COMPATIBILITY
MOTOR FRAME SIZE
BLOWER MFG
Aerovent
Buffalo
FAN SIZE
33
36
38
33
38A
43B
MAX MOTOR
FRAME SIZE
256 T
324 TC
324 TC
286 C
286 C
365 C
T - Foot mounted
C - Flanged
TC - "C" Flanged Foot mounted
MOTOR SIZES
MFG
Sabina Electric
Randtronics and
Rel iance
Wer
Eaton
MOTOR
TYPE
DC
DC
EC
EC
MOTOR
RATING
(HP)
20
25
30
20
25
30
20
25
30
20
25
30
MOTOR FRAME SIZE
DRIP-PROOF TOTALLY ENCL.
284 A 324 A
284 A 328 AT
324 A 366 AT
259 AT 327 AT
287 AT 328 AT
288 AT 366 AT
286 TD
324 TSD
324 TSD
286 T 286 T
284 T 284 T
286 T 286 T
6-24
-------
service factor of 1.15 at 40.5°C (105°F), and noncorrosive
air is being passed over the motor.
In conclusion, there are no problems in selecting
a direct-drive, variable-speed motor for use with a centri-
fugal blower. In contrast, variable-speed motors are avail-
able for use with vaneaxial blowers, but careful examination
of design features must be carried out to ensure that the
selected motor and blower are compatible. If a totally
enclosed motor is required for use with a vaneaxial blower,
an eddy current motor appears to offer some size advantages
over a DC unit.
6.5.9 Size and Weight
As discussed in the Task 5 report, there is a
substantial difference in the weight and the dimensions of
the two blowers being considered for this application.
These differences necessitate different blower mounting
techniques and ducting designs.
Table 6-4 provided data on the weight of each
style blower. As was seen from this data, the mass of the
vaneaxial units, as manufactured by Aerovent and Buffalo
Forge, is one-fifth to one-half the mass of comparable
centrifugal units. These masses do not include the motor.
The mass of eddy current and DC motors, comparably
rated, are similar. Typical motor weights are shown in
Table 6-8. The motor mass is of the same magnitude of the
weight of the vaneaxial units and one-third to one-half the
weight of the centrifugal units.
6-25
-------
Table 6-8. MOTOR WEIGHTS
MANUFACTURER
Rel iance
Eaton
MOTOR TYPE
20 HP DC Drip-Proof
25 and 30 HP DC Drip-Proof
20 and 25 HP DC Totally
End .
30 HP DC Totally Encl .
20 HP EC Drip-Proof
25 HP EC Drip-Proof
30 HP EC Drip-Proof
FRAME
SIZE
259 AT
287 AT
327 AT
366 AT
286 T
284 T
286 T
WEIGHT
(kg)
188
261
351
454
252
252
454
As a result of the centrifugal unit's size and
mass, it must be securely mounted to the test cell floor.
Therefore, in order to accomodate various-sized motorcycles,
while maintaining a fixed distance between duct exit and the
front wheel of the bike, telescoping ducts must be used.
Mounting configurations for the vaneaxial unit are
more flexible because of its lighter weight and smaller
size. Conversations with vendors have confimred that the
vaneaxial system can be mounted on rails and be moved to the
appropriate position for testing the various sized units.
6.5.10 Flow Conditioning
The contract specifies that the outlet flow from
the blower shall be uniform to ±20 percent across the outlet
area as measured at the center of the area as compared to
the center of each quarter area. An analysis of the velocity
profiles, as it impacts on the ducting design, will be
examined in Section 6.7 but consideration should be given at
this time to the general characteristics of each style
blower and available flow-conditioning accesories.
A propeller or axial flow-type wheel imparts a
rotating or screw action to the exit air. In a vaneaxial
6-26
-------
fan, stationary guide vanes are located directly behind the
fan wheel. These vanes reduce air turbulence and partially
convert the rotative energy imparted to the air to axial
energy. The result is an increase in efficiency, a reduc-
tion in noise and nonswirling air flow leaving the fan. In
addition, an inlet bell should be installed on the entrance
to the vaneaxial unit. This reduces the static pressure
loss associated with introducing the air into the fan.
The velocity profile of the air exiting a centri-
fugal blower is not uniform across the entire cross section
of the blower exit. This nonuniformity results from the
centrifugal energy imparted to the air by the wheel. The
degree of nonuniformity will be a function of the impeller
speed with maximum velocity asymmetry occuring at maximum
rotational speeds. Flow conditioning techniques, such as
turbulators and vane control can be used to condition the
exit air. However, the exact flow-conditioning design
cannot be developed until more specific data is available.
6.5.11 Costs
Typical blower costs were shown in Table 6-4. As
shown, vaneaxial units are approximately 50 percent less
costly than comparable centrifugal units. In addition, the
costs associated with shipping the vaneaxial units are also
less because of its lower weight.
6.6 EVALUATION CRITERIA - MOTOR CONTROL AND SYSTEM
PERFORMANCE
6.6.1 Accuracy
As specified in the contract, the linear air
velocity from the blower shall be within 10 percent of the
6-27
-------
motorcycle speed for speeds between 10 and 100 kph and
within ±1 kph at speeds below 10 kph. In order to achieve
this goal, a variable speed drive equipped with a tachometer
follower control or shaft coupling circuitry is recommended.
This control enables the driver to follow a signal from the
tachometer generator located on the dynamometer roll. The
accuracy of the control is dependent on the design of the
control circuitry, the characteristics of the tachometers
located on the dynamometer roll and fan wheel shaft, and the
linearity of the relationship describing blower output
versus fan wheel rpm.
As described in Section 6.5.1, the vaneaxial
blower exhibits a linear relationship between blower output
and impeller rpm. Centrifugal blowers do not have this
characteristic. Therefore, with the recommended circuitry,
the accuracy of a centrifugal blower system will be less
than observed with the vaneaxial system.
The manufacturers of variable speed motors and
controllers define the accuracy of their system in terms of
speed regulation. Regulation is defined in terms of a
percentage of maximum speed. Table 6-9 summarizes the
characteristics of representative drives. In some cases,
the manufacturer has a basic specification for speed regula-
tion and optional circuitry for increased regulation. These
options have also been identified in the table.
6-28
-------
Table 6-9. MOTOR SPECIFICATIONS
MANUFACTURER
Eaton
Louis Allis
WER
Sabina
Randtronics
MOTOR
TYPE
EC
DC
EC
DC
EC
DC
DC
DC
SPEED REGU-
LATION (%)
1
2, 1
1, 0.5, 0.1
2, 1
2, 1, 0.1
2, 1. 0.1
3, 1
2, 1, 0.5,
25
SPEED
RANGE
1695-50
1695-88
1695-50
1695-50
1710-50
1750-17
1750-17
1750-0
RATED TORQUE (NT-M)
20 HP
33.9
-
-
79.0
-
25 HP
33.9
-
-
98.8
-
30 HP
44.0
-
-
118.5
-
In order to achieve the accuracy goal defined in
the contract, at least 1 percent speed regulation must be
obtained. (This assumes that the blower output is perfectly
linear.) As shown in Table 6-9, this degree of regulation
is available from several suppliers, using either eddy
current or DC systems. The methods used to achieve this
degree of control varies from manufacturer to manufacturer.
Some utilize tachometer feedback control to achieve this
regulation, while others use proprietary techniques to
improve the system's speed regulation characteristics.
In conclusion, from the standpoint of speed regula-
tion (i.e., accuracy), the variable-speed drive coupled to a
vaneaxial blower is superior to a system employing a cen-
trifugal blower.
6.6.2
Speed Range
The speed range required to satisfy the goals
outlined in the contract is 100-to-l or 1,750-17 rpm. This
6-29
-------
is equivalent to a wind velocity range of 100 kph to 1 kph.
Table 6-9 details the speed range of the various drives.
A motor's speed range is defined at 100 percent
torque, continuous duty, and torque is defined by equation (6-5)
r = I 0< (6-5)
where
I is the rotational inertia and o< is the angular accelera-
tion. The required angular acceleration rate is controlled
by the driving cycle and the air delivery characteristics of
the fan wheel. The maximum angular acceleration is defined
by equation (6-6)
Tmax _ (max cycle acceleration)(blower exit area) (6-6)
(blower output per revolution)
Maximum torque requirements have been calculated for a
representative vaneaxial unit (Aerovent 361) and centrigual
blower (New York Blower 449). These are listed below.
Aerovent 361 rmax = 27.1 nt.-m (240 Ib-in)
New York Blower 449 rmax =125.6 nt.-m (1,112 Ib-in)
Since maximum acceleration rates in the FTP occur at low
speeds, the maximum torque demand will also occur at low
motor speeds. Typical torque levels available were shown in
Table 6-9. It should be noted that the eddy current motor
is a torque transmitter in contrast to a DC motor which is a
torque converter. As a result, the eddy current drive
provides an essentially-constant torque over its entire
speed range rather than torque being a function of rpm.
Two conclusions are readily apparent after review-
ing the data in Table 6-9. First, eddy current drives
6-30
-------
cannot be operated at low enough rotational speeds to
satisfy the 100-to-l operating range requirement. These
units could be operated at lower speeds, but overheating
might occur and auxiliary cooling would require additional
space and add cost and complexity. DC systems do have
sufficient dynamic range to fulfill this application.
Secondly, the drives under consideration,.based on
power requirements, cannot provide sufficient torque to
drive a centrifugal blower at the lower speed and torque
imposed by the FTP. The motor can be oversized to provide
sufficient torque. It is estimated that a 30 kw (40 hp)
motor would be required to drive a centrifugal blower versus
15 kw (20 hp) motor to drive a vaneaxial unit.
6.6.3 Response Time
The control system which is selected for this
application must be capable of tracking the motorcycle
speed; both its accelerations and decelerations. As
discussed in Section 6.5.2, the motors under consideration
are capable of accelerating the blower. However, without
external braking, the decelerations cannot be followed.
Eddy current systems are available with two differ-
ent types of brakes. A friction brake can be added, but
this type of unit has been designed to bring the shaft to a
stop. It is not intended to slow the rotation of the unit
in a controlled manner. In order to accomplish the controlled
deceleration of the unit, an eddy current or electromagnetic
braking system must be added. Sometimes referred to as
Mutatrol, this option is available from all three known
suppliers of eddy current motors, and consists of a second
eddy current unit which performs the function of a power
absorption unit or brake. Its braking capacity is compar-
able to the accelerating capacity of the eddy current motor.
The cost impact of this device is examined in Section 6.6.9.
6-31
-------
DC motors are also available with external braking
packages. These auxiliary braking systems, however, are
friction brakes which cannot accurately control decelerations.
Controlled decelerations with a DC system can only be achieved
with a regenerative DC system. These systems are effectively
compact motor-generator sets.
As described by WER Industrial, a DC motor with
fixed-field excitation carries an armature current which
will produce torque propotional to the armature current and
in the direction of the armature current. In the regenera-
tive unit, the armature current can flow in either direction
under controlled conditions, resulting in what is termed as
four quadrant drive operation. In order to accomplish this,
two SCR modules are placed back-to-back across the DC motor.
In the motoring mode, a module will act as a rectifier to
convert the incoming AC power to DC power for the motor
armature. When called upon to slow the motor, the other
module will function and operate as an AC line commutated
inverter. By phasing on the SCRs at the proper phase angle,
this module will convert the DC armature voltage to an AC
wave form so that current is allowed to flow through the
motor armature back into the AC line. In order to prevent
the two modules from operating simultaneously, the controller
should contain "lock out" or "dead band" circuitry. The
regenerative DC system is ideally designed for this applica-
tion, and it is the DC counterpart to Mutatrol.
As discussed in the previous section, many of
these motors may have insufficient torque to accelerate a
centrifugal through the expected short-term accelerations.
Since the expected decelerations are of the same magnitude
as the acclerations, some difficulties may be encountered in
decelerating a centrifugal blower, with either a regenera-
tive DC system or eddy current unit.
6-32
-------
6.6.4 Delivery
Inquires were made to three manufacturers «of eddy
current units and seven manufacturers of DC and regenerative
DC systems regarding estimates of delivery. Estimates
ranged from 3 weeks to 26 weeks for all three motor styles.
The manufacturers advised that delivery schedules fluctuate
wildly and are usually dependent on the availability of the
motor. These schedules appear to be reasonable and in line
with the delivery of the other dynamometer components.
6.6.5 Size and Weight
Typically, the eddy current system, DC system and
the regenerative DC system consist of three components.
These are the motor, the controller, and the operator's
station.
Motor sizes were examined in Section 6.5.8 and
comparable motor sizes were shown in Table 6-7 and 6-8. As
previously discussed neither style of motor offers any
significant weight or size advantage.
Figure 6-7 shows the typical dimensions of the
controller and operator's station. The controller is typic-
ally wall-mounted while the operator's station is designed
for hand-held operation. As shown, the eddy current
controller is somewhat smaller than its DC counterpart.
However, the size of the DC controller is not overly large,
nor does it limit the usefulness of the DC system.
6.6.6 Adaptability
The use of an eddy current motor or DC motor in
conjunction with tachometer follower circuitry and controlled
braking capability, results 1n a versatile system capable of
6-33
-------
B
OFF
ON
O
CONTROLLER
OPERATORS STATION
MANUFACTURER
Eaton
Sabi na
Re 1 iance
MOTOR
TYPE
E.G.
o.c.
D.C.
A
30.5
91.4
97.8
OIMENS
B
24.1
76.2
67.3
IONS (CN
C
15.3
30.5
38.7
4)
D
24.1
24.1
26.1
E
9.1
7.0
8.6
FIGURE 6-7. CONTROLLER AND OPERATOR STATION DIMENSIONS
6-34
-------
tracking motorcycle speeds. The limit to this tracking
capability is not the controller, but rather the motor's
torque/power limitations. As discussed in Section 6.5.2,
less torque is required to accelerate a vaneaxial unit than
a centrifugal blower. Therefore, should it be necessary to
follow faster accelerations than found in the LA-4, the
vaneaxial systems would be most adaptable.
In addition, most manufacturers supply tachometer
follower circuitry which can be manually overidden. This
would permit the blower to be operated at speeds which
differ from the motorcycle speed. In this mode, studies
could be conducted to examine the effects cooling has on
motorcycle emissions.
Finally, the versatility of a nonregenerative DC
motor is restricted because of its limited braking capacity.
6.6.7 Reliabi1ity/Maintenance
The eddy current motor requires less maintenance
than a DC drive. The simple mechanical design of the eddy
current drive eliminates brushes, commutators and slip-
rings; all of which are used in DC drives and require peri-
odic maintenance. In an eddy current drive, the only mechan-
ical item which may require maintenance are the bearings
between the input and output members of the magnetic clutch.
The typical life of these bearings is in excess of 50,000 hours
The superior reliability of an eddy current unit
is supported by data from Louis Allis, a manufacturer of
eddy current, DC, regenerative DC and variable-speed AC
drives. In May 1975, their warranty expenses for eddy
current drives were 0.05 percent of sales versus 0.7 percent
for all types of DC systems and 4.2 percent for variable AC
uni ts.
The controllers used for both styles of drives
consist of sophisticated SCR/solid-state circuitry, which is
known for reliability and ease of maintenance. Periodic
6-35
-------
maintenance consists basically of periodic inspection and
cleaning of the controller.
All manufacturers recommend the acquisition of a
spare parts kit. These consist of spare fuses and circuit
boards.
6.6.8 Interferences
The system utilized in this application should not
be affected by other electrical and environmental elements
in the test area. Temperature, and humidity will not be a
factor since these are controlled in the test cell. Alti-
tude does effect the performance of the blower and driver,
but the altitude of Ann Arbor, Michigan, is not sufficient
to pose any problems. At altitudes over 900m, modifications
of blower size and motor rating must be considered, if sea
level testing is to be simulated.
Electrical interference may be a problem. Kawasaki
in Shakope, Minnesota, has experienced some problem with
their DC system as a result of RF interferences. Consider-
ation should be given to interference associated with high-
energy capacitor discharge ignition systems. In addition,
manufacturers recommend the use of isolation transformers to
eliminate the effects of power-line fluctuations and warn
the user to examine grounding requirements and connections.
Interference effects are difficult to foretell and
are affected by the specific design of the drives. At this
time, none of the drives examined appear advantageous in
this respect. However, the specification should address
this problem.
6.6.9 Costs
Representative prices have been obtained from
several manufacturers of eddy current drives and regenerative
6-36
-------
DC drives. Prices for DC drives have not been determined
because they lack controlled deceleration capabilities. The
system prices are all F.O.B. manufacturer's plant and include,
as a minimum, the motor, controller with 1 percent speed
regulation, tachometer follower circuitry, controlled braking
capability and isolation transformer. Also, all prices are
for open-type motors and do not reflect any O.E.M. discounts
or price premiums for speedier delivery. Prices are shown
in Table 6-10.
Table 6-10. MOTOR/CONTROLLER COSTS
MANUFACTURER
Sabina
WER Industrial
Louis Allis
MOTOR
TYPE
RDC
EC
RDC
EC
RDC
SYSTEM
15 KW 19 KW
(20 HP) (25 HP)
7561 8190
8655
9109
6147 6147
7180 8245
COST
22 KW
(30 HP)
8751
-
OEM
DISCOUNT
28%
10%
10%
10%
The costs of the two systems are comparable. The
price spread among manufacturers appears to be greater than
any spread between types of drive.
6.7
OUTLET DUCT DESIGN
The contract and NPRM define the geometrical
configuration of the cooling air outlet duct. These have
been summarized in Section 6.2 of this report. In this
section, these specifications will be analyzed with respect
to their impact on the simulation of motorcycle cooling.
The intended purpose of the cooling system is to
simulate cooling conditions encountered in road operation
6-37
-------
and thereby control motorcycle engine temperatures to within
±10 percent of those observed on the road. With unlimited
funds, this could easily be achieved by placing the motor-
cycle in a wind tunnel of sufficient magnitude. The costs
to accomplish this would be substantially greater than the
already high costs estimated in Sections 6.5 and 6.6. By
properly configuring the outlet duct, it is believed that
the simulation of road-cooling condition can be achieved
within a realistic cost constraints. This will be verified
during Phase II of this program.
6.7.1 Outlet Configuration of Ducting
As specified in the contract, the outlet of the
cooling duct shall have a rectangular cross-section of
2
0.5m , and the bottom of the ducting shall be located about
150mm (5.91 inches) above dynamometer level. These dimen-
sions resulted from the physical measurement of the engine
configuration of various motorcycles.
2
The 0.5m rectangular cross-sectional area was
selected because the largest engine measured could be placed
within it. If the engine were placed within a duct of this
size and configuration, cooling simulation would be achieved.
Unfortunately, the ducting will not encompass the engine,
although the air exiting the ducting will be directed toward
the engine. As shown in Figure 6-8, the velocity of the air
discharged decreases at increasing distances from the duct
exit. Additionally, there is substantial broadening of the
discharge pattern as the distances from the exit increase.
This velocity pattern will be further disrupted and
broadened because the air must pass around the front wheel
of the motorcycle and portions of the motorcycle frame
before reaching the engine. Broadening would not be so
severe during road operation because the air flow about the
front wheel would be contained by the air streamlines, some
distance from the side of the engine.
6-38
-------
I
CO
1C
15 20
LENGTH—NOZZLE DIAMETERS. D
:
,
35
FIGURE 6-8. AIR DISCHARGE PATTERN
-------
D
PIGURE 6-9. COOLING SYSTEM LAYOUT
6-40
-------
The broadening effect can be minimized by design-
ing the outlet duct to straddle or envelope part of the
front wheel of the vehicle. This is shown in Figure 6-9.
The effects this would have on engine-cooling are difficult
to quantify because of the complex nature of the heat trans-
fer between the engine and the air, and the various geometries
of the motorcycle engine. It is expected that these effects
will be analyzed experimentally during Phase II of this
program.
6.7.2 Velocity Profiles
The contract defines the uniformity of the velocity
profile of the air exiting the outlet duct. Figure 6-10
displays fully-developed velocity profiles for laminar and
turbulent flow. As a result of the 100-to-l range of air
velocities expected, both laminar and turbulent conditions
are expected. The ducting lengths will be short, however,
and fully-developed flow conditions will not occur.
Figure 6-11 shows the laminar velocity profiles in
a start-up condition as shown. Nearly flat velocity profiles
occur as duct length decreases. The development of turbulent
flow profiles is similar. Therefore, if the length of the
outlet duct is less than three times the fan diameter, the
velocity uniformity specification can be attained.
6.8 CONCLUSIONS
Table 6-11 summarizes the evaluation study. As
quantified in the summary table, a vaneaxial system is
superior to a system employing a centrifugal blower. The
vaneaxial blower has distinct performance advantages result-
ing from its smaller physical size, characteristics, and
lower cost.
6-41
-------
Tube center
1.0
0
1.0 0.8 0.6 0.4 0.2
1.0
0.2
0
0.2 0.4 0.6 0.8 1.0
r/R *
FIGURE 6-10. QUALITATIVE COMPARISON OF LAMINAR AND
TURBULENT VELOCITY DISTRIBUTIONS
6-42
-------
max
Tube center
1.0 0.8 0.6 0.4 0.2
I 0.2 0.4 0.6 0.8 1.0
r/K—^
FIGURE 6-11. VELOCITY DISTRIBUTION FOR UNSTEADY STATE
"START-UP" FLOW IN A CIRCULAR TUBE
6-43
-------
Table 6-11. WEIGHTED RATINGS
CRITERION
I. Blower Performance
1. Flow as a funtion of
RPM and AP
2. Rotational Inertia
3. Delivery
4. Noise
5. Reliability/Maintenance
6. Adaptability
7. Power Consumption/Eff .
8. Motor Compatibility
9. Size and Weight
10. Flow Conditioning
11. Cost
II. Motor/System Performance
1 . Accuracy
2. Range
3. Response Time
4. Delivery
5. Size and Weight
6. Adaptability
7. Reliability/Maintenance
8. Interferences
9. Costs
Total Rating
Ranking
SYSTEM NO. (SECTION 2.2)
Vaneaxial Blower
EC
5.0
2.7
2.4
2.4
3.2
2.4
1.4
2.1
2.4
2.0
2.0
4.0
1.8
2.7
2.4
2.1
2.8
2.8
2.1
2.1
50.8
2
DC
5.0
2.7
2.4
2.4
3.2
2.4
2.1
1.4
2.4
2.0
2.0
1.0
3.6
0.9
2.4
2.1
0.7
2.1
2.1
2.1
45.0
3
RDC
5.0
2.7
2.4
2.4
3.2
2.4
2.1
1.4
2.4
2.0
2.0
4.0
3.6
2.7
2.4
2.1
2.8
2.1
2.1
2.1
51.9
1
Centrifugal Blower
EC
5.0
1.8
2.4
3.2
2.4
2.4
1.4
2.8
1.8
1.0
1.5
2.0
0.9
1.8
2.4
2.1
1.4
2.8
2.1
2.1
40.3
5
DC
2.0
1.8
2.4
3.2
2.4
2.4
2.8
2.8
1.8
1.0
1.5
1.0
0.9
0.9
2.4
2.1
0.7
2.1
2.1
2.1
38.4
6
(
-------
The evaluation also indicates the regenerative DC
drives and eddy current motor/brake systems are suitable for
this application. The first of these options can meet the
100-to-l dynamic range requirement but it is felt that a 30-
to-1 range of performance should be acceptable for this
application. Assuming this smaller range is acceptable,
then the eddy current system offers slight size, maintenance
and cost advantages. A nonregenerative DC drive is unaccept-
able because it would be unable to follow motorcycle
decelerations.
6-45
-------
Section 7
TASK 7 - DYNAMOMETER SYSTEM ANALYSIS
7.1 INTRODUCTION
In this report a cost-effectiveness study comparing
three alternative dynamometer designs will be presented.
The performance characteristics of these alternatives will
be compared to the performance requirements developed in
previous task reports and detailed in the Notice, of Proposed
Rulemaking (NPRM) and the contract. As a result of this
study and evaluation, a design will be recommended. A
specification for this recommended system will, subsequently,
be developed.
7.2 REVIEW OF REQUIREMENTS
Paragraph 85.478-15 of NRPM defines the inertia
and road-load requirements for a motorcycle dynamometer.
These requirements are more explicitly defined in the contract
and the Task 2 Report.
As stated in the contract, the dynamometer should
be capable of simulating inertia in the range from 100 kg
(220 Ibs) to 700 kg (1,543 Ibs). The simulation should
provide discrete inertia steps with intervals of 10 kg
(22 Ibs).
Road-load data on ten motorcycles was provided by
the Project Officer to Olson Laboratories for analysis. The
analysis was presented in Task 2. It indicated that at
7-1
-------
motorcycle speeds of 110 kph, which was the maximum speed
encountered during the durability cycle, a maximum power
absorption capacity of 14.7 kw (20 HPm) was required. This
assumes that flywheels will be used for inertia simulation.
If inertia is to be simulated electrically, the absorption
capacity would increase to approximately 37 kw (50 HPm), and
an additional 18.4 kw (25 HPm) motoring capacity would be
needed.
As discussed in the Task 2 Report, the road-load
characteristics of various motorcycles differ substantially,
and one characteristic curve is insufficient to simulate
road-load condition-s of all motorcycles. This can be achieved
by using a controller which can accept the algorithm developed
by the EPA and which loads the power absorption unit (PALI),
as required. These requirements must be satisfied by the
alternative selected.
7.3 REVIEW OF RECOMMENDATIONS FROM TASKS 2, 3, AND 4
In Tasks 2, 3, and 4, evaluations of PALI roller
configurations and inertia simulation techniques were pre-
sented, and recommendations were made. These findings were
discussed in detail at a review meeting held with the Project
Officer and other members of the EPA staff on May 21, 1975.
Subsequently, in a letter dated June 18, 1975, the Project
Officer directed Olson to examine three systems. The three
systems are:
1. Eddy current PAU - inertia simulation by
mechanical (flywheel) means.
2. DC or AC motor generator - inertia simulation
by mechanical (flywheel) means.
7-2
-------
3. DC or AC motor generator - partial mechanical
plus electrical simulation using the motor
generator.
Additionally, each of the three systems would employ a
smooth, single dynamometer roll having a diameter of 530.5 mm
Olson Laboratories concurs that these alternatives
are viable approaches which would result in systems which
achieve the design goals outlined in Section 7.2.
7.4 DESIGN ALTERNATIVES
As stated in Section 7.3, the Project Officer has
requested that the three design alternatives be examined in
detail. As a result of this study, a final design will be
developed and specifications will be formulated. In this
section, evaluation criteria will be defined. Subsequently,
a cost-effectiveness study of the design alternatives will
be conducted.
In reality, there are five systems which should be
considered. These combine the three types of power absorbers
with the two methods of inertia simulation. The five combina^
tions are:
1. DC motor generator - inertia simulation by
mechanical (flywheel) means.
2. DC motor generator - inertia simulation by
partial mechanical plus electrical techniques.
3. Eddy current PALI - inertia simulation by
mechanical (flywheel) means.
7-3
-------
4. AC motor generator - inertia simulation by
mechanical (flywheel) means.
5. AC motor generator - inertia simulation by
partial mechanical, plus electrical techniques
In the Task 2 analysis, the characteristics of the
AC motor generators were examined. As reported, adjustable-
speed AC drives are advantageous in applications; wherein
sparkless operation, ultra high speed, multiple-drive synchro-
nization, and/or adverse environmental conditions are a
factor. None of th-ese factors seem to be considerations in
this application. Additionally, Louis Allis, a manufacturer
of both AC and DC systems, reports that costs of an AC
system are twice that of a comparable DC system. Therefore,
further considerations of AC systems seem unwarranted.
Accordingly, this evaluation will be limited to those alterna-
tives which use eddy current or DC PAU's.
7.4.1 Evaluation Criteria and Ranking System
In order to arrive at the most cost-effective
design for a motorcycle dynamometer, the various combinations
of system components must be systematically and comprehen-
sively evaluated and compared. The analytical procedures
must consider all factors and characteristics that contribute
to system performance and costs, influence system operational
procedures, and affect system maintenance, maintainability,
and reliability.
All of the alternatives will be evaluated with
respect to each of the criteria and will be rated on a scale
of l-to-5; where 1 is unacceptable, 3 is acceptable, and 5
is outstanding. Additionally, a weighting factor has been
designated for each of the criteria, defining its relative
importance. Each of the ratings will be adjusted by using
7-4
-------
the weighting factor, and the rankings of the alternatives
will be calculated by summing the weighted ratings.
The evaluation criteria, their respective weighting
factors and the resultant ratings are listed in Table 7-1.
7.5 EVALUATION OF DYNAMOMETER SYSTEMS
7.5.1 DC PAU - Inertia Simulation by Partial Mechanical
Plus Electrical Techniques
Performance
In this configuration the system would consist of
a regenerative-type DC motor which functions as the PAU, a
controller, and a single flywheel having an inertia of
300 kg, midway between the 100 kg and 700 kg limits, specified
in the contract. Prior to the test, the operator would
input into the controller the mass of the motorcycle, and
the constants of an algorithm which define the road-load
characteristics of the vehicle being tested.
During the test, the controller would provide
signals to the DC motor based upon the input signals to the
controller and the characteristic data inputed to the
controller prior to the test. The continuous inputs which
are fed to the controller are load, as measured by a load-
sensing element or torque bridge, and vehicle velocity, as
measured by a tach generator or optical encoder mounted on
the shaft of the dynamometer roll. The controller would
calculate the road-load from the algorithm and compare it to
the load as measured by the load cell. The resulting error
signal would be used to provide a signal to the DC motor,
which would result in additional power absorption or motoring.
It is estimated that 100 milliseconds would be required to
perform this comparison and generate the error signal. This
7-5
-------
Table 7-1. SYSTEM ANALYSIS FACTORS
FACTORS
I. Performance
1. Ability to Simulate Road
Conditions
2. Ability to Simulate Vehicle
Inertia
3. Ability to Accommodate Various
Morotcycles
4. Fail -Safe Provisions
5. Ability to Simulate Durability
Cycle
6. System Repeatability
II. Operation
1. Operating Personnel Training
2. Pre-Test, Set-Up Procedures
3. Supporting Utilities Required
4. Test Data Acquisition
5. Facility Spatial Requirements
III. Maintenance
1. System Complexity
2. Maintainability of System
3. Reliability of Components
4. Preventative Maintenance
IV. Costs
1. Purchase Price
2. Installation Costs
3. Annual Operating Costs
Cumulative Ratings
Ranking
WT.
FACTOR
1.0
1.0
1.0
1.0
0.8
1.0
0.5
0.6
0.6
0.4
0.5
0.6
0.8
0.8
0.6
0.8
0.5
0.6
DYNAMOMETER CONFIGURATION RATING
DC System
Elect. Inertia
Unweighted
2
2
3
3
3
2
3
2
4
3
4
3
3
3
3
2
3
3
Weighted
2.0
2.0
3.0
3.0
2.4
2.0
1.5
1.2
2.4
1.2
2.0
1.8
2.4
2.4
1.8
1.6
1.5
2.4
38.4
3
DC System
Mech. Inertia
Unweighted
4
4
3
3
3
3
3
3
3
3
2
3
3
3
3
1
3
3
Weighted
4.0
4.0
3.0
3.0
2.4
3.0
1.5
1.8
1.8
1.2
1.0
1.8
2.4
2.4
1.8
0.8
1.5
2.4
41.6
2
Eddy Current
Mech. Inertia
Unweighted
4
4
3
3
4
3
3
3
2
3
2
4
3
3
3
3
3
3
Weighted
4.0
4.0
3.0
3.0
3.2
3.0
1.5
1.8
1.2
1.2
1.0
2.4
2.4
2.4
1.8
2.4
1.5
2.4
45.0
1
-------
time is based on a 10-mi11isecond computation time and a
100-mi11isecond period for sensing and averaging the load
signal. The picking and averaging time would occur concur-
rently with the calculation time. The system response time,
defined as the elapsed time between signal input to the
controller and actual armature current change resulting in
PAU load change, would be on the order of 500 to 1,000
milliseconds. This response time is controlled by the
inertia of the motor armature and the controller design,
which would vary with motor design.
By utilizing a flywheel having an equivalent
inertia approximately at the mid-point of the inertia range,
the motor will only be required to simulate half of the
total inertia spread. For vehicles having inertias greater
than the mid-point value, the DC regenerative motor would
operate in a regeneration (absorption) mode; thus, adding to
the simulated inertia seen by the motorcycle. For those
motorcycles with an inertia mass less than that of the fixed
flywheel, the DC motor would function in both motoring and
absorption modes. As a result of having to function in both
modes, response times may be increased, and road-load simula-
tions for the smaller vehicles may not be as accurate as the
simulation for the larger vehicles. In order to have a
transition from a motoring mode to an absorption mode, the
system response time cited above may be increased by an
additional 500 milliseconds. This added time will be affected
by the inertia of the motor's armature and the characteristics
of the controller.
The addition of more than one flywheel, decreases
the capacity for total inertia simulation upon the motor.
Thus, the motor will work more in a regenerative mode than
in a motoring mode. This will effectively reduce the size
requirement of the original motor selection, but, concurrently,
it will require the addition of flywheels, bearings, sub-
frame, clutching and de-clutching mechanisms. The cost for
7-7
-------
these mechanisms must be balanced against the differential
cost of a smaller DC regenerative motor.
In this application, repeatability of operation is
critical. The degree of repeatability will be affected by
the stability and repeatability of the road-load and velocity
measurements and the controller. The use of the load cells
are recommended because they are accurate, rugged and unaf-
fected by temperature fluctuations. Armature current measure-
ments can be used to determine motor load, but the accuracy
of these measurements are affected by motor temperature. It
has been estimated by Sabina Electric, a manufacturer of
regenerative DC sys-tems, that repeatability within 0.1 per-
cent can be achieved utilizing an optical encoder for measur-
ing dynamometer roll speeds and a temperature-compensated
load cell for power measurements.
Fail-safe provisions are standard in most DC
systems. Accordingly, the DC regenerative motor system is
typically equipped with an auxiliary cooling fan that is
connected to the motor and force-ventilates the motor, but
does not add significant inertia to the system. Additionally,
the controller incorporates an overload system which allows
for low speed, high torque operation without the danger of
overload, and is equipped with an armature current fuse
protection. This latter has an adjustable set point and
limits the current to a specific level. Typically, it can
be set between 125 and 200 percent maximum current rating.
The controller also has an electronic overload switch which
trips out the motor when it sees an instantaneous overload
in excess of the set point. Additional protection is provided
against excessive acceleration or deceleration.
In summary, with respect to performance, the DC
system, utilizing electrical techniques to simulate inertia,
can approximately simulate road conditions. The simulation
will be limited by the response-time characteristics of the
7-8
-------
system which is affected by motor design.
By properly sizing the regenerative system, the
dynamometer can be used for both durability and certification
testing, and its use will not be affected by the size or
performance characteristics of commercially-available "street"
motorcycles.
Operation
The controller for this system will be quite
complex, but the unit itself can be operated by a trained
technician. The level of training will probably be more
extensive than that required to operate a Clayton hydrokinetic
unit and will probably be less than the training required to
operate a programmable, automatic light-duty vehicle dura-
bility dynamometer such as produced by Labeco.
Pre-test set-up procedures will not be extensive.
They will basically consist of placing the vehicle on the
dynamometer, restraining it, and dialing into the controller
the appropriate inertia setting and road-load algorithm
constants. Since the unit requires no utilities except
power, typically 460 VAC, 60 Hz, 3 0, no other connections
have to be made.
One pre-test procedure which must be periodically
performed is the calibration of the dynamometer. Calibration
procedures for dynamometers equipped with flywheels are
straightforward and well defined. However, procedures for
verifying the inertia simulation accuracy of this system
have not been fully defined. These procedures, when developed,
may be quite complex, requiring additional manpower and
equipment.
As in all the sytems being examined, a driver's
aid must be provided. At a minimum, it will provide data on
instantaneous speed and horsepower. Additional readouts
such as armature current may be useful, but none are mandatory.
7-9
-------
Of the three configurations being examined, the DC
system with electrical inertia simulation will occupy the
least amount of space. Although the DC motor is larger than
an eddy current brake, the lack of a flywheel system results
in a net decrease in space requirements. Of course, the pit
depth and width will be a function of roller length and DC
motor frame length. The latter will vary from manufacturer
to manufacturer.
Maintenance
The system proposed here is complex. A DC regenera^
tive system with SCR-type controller, will consist of a
velocity signal generator mounted on the roll, a load cell,
and an operational box which will use the voltage signal
from the load cell to provide the proper force signal to
load the PAD. This operational box can be in the form of a
mini-computer or microprocessor.
The components of this DC system are all designed
as standard components but not as a total system. There are
no electronic components that are specifically designed for
this application. All operational amplifiers are standard
industry type. All components are shelf items with the
exception of enclosures, etc.
Preventive maintenance is required to maintain the
excellent performance characteristics of a DC motor. Items
requiring periodic maintenance include brushes, commutator
and stator. Brushes should be replaced every 5,000 hours of
operation. Of course, undersized or oversized brushes will
have a detrimental effect on the life of the brus'h. The
brush selection is based upon the application and must be
determined in the intitial design and tested in the field
after the first or second brush change-out.
With the exception of the replaceable electrical
motor components, a realistic life of the system is 10 years
7-10
-------
with recommended maintenance on the motor. Recommended
maintenance includes motor teardown every 3 years. This
includes cleanout, bearing replacement and commutator dress
up. Periodic brush replacement should occur approximately
every 3 months. With the exception of human error or mechanic-
ally caused problems in the electronics, reliability of the
electronic components should be high. Since most of the
electronics are self-compensated for drift component, age is
not a problem.
Costs
Tables 7-2 and 7-3 summarizes the prices of various
DC and eddy current dynamometers. These prices were provided
by Meidensha, Ltd., and provide valuable information regarding
the relative costs of the alternative configurations. A
comparable estimate was received from Consine Dynamics, Ltd.
As shown in Table 7-2, the DC system with electrical inertia
simulation is the least expensive of all the configurations
being examined. However, as shown in Table 7-3, the cost
impact of enlarging the power absorption capacity to fulfill
the requirements of the durability cycle is significant,
because the capacity of the regenerative system must be
doubled. The installation costs of the three systems are
comparable.
Typically, the components described above are
available within 3 to 6 months after an order is placed.
7.5.2 DC System with Mechanical Inertia Simulation
Performance
This system is configured similarly to the DC
system utilizing electrical inertia simulation. The basic
difference is the addition of six flywheels having equivalent
inertias of 10, 20, 40, 80, 160, and 320 kg.
7-11
-------
Table 7-2. PRICE AND SPECIFICATION FOR EMISSION DYNAMOMETER
(Al) D.C. Chassis Dynamometer
with Flywheels
Total Price: $48,150.00
(A2) D.C. Chassis Dynamometer
with two Flywheel and
70kg Electrical Intertia
Simulation
Total Price: $38,950.00
(A3) D.C. Chassis Dynamometer
with one Flywheel and
150kg Electrical Inertia
Simulation
Total Price: $33,950.00
(A4) D.C. Chassis Dynamometer
with 300kg Electrical
Inertia Simulation
Total Price: $31,270.00
(A5) Eddy Current Chassis
Dynamometer with
Flywheel only
Total Price: $38,950.00
ROLLER FLYWHEEL
Steel made, 10kg+20kg+40kg
530.5mm +80kg+150kg
direct control
$3,650.00 $20,300.00
80kg+150kg
" direct control
$3,650.00 $10,000.00
150kg
remote control
H
$3,650.00 $4,600.00
Fixed Flywheel
180kg including
roller and dyna-
mometer inertia
mounted on roller
shaft
$3,650.00 $770.00
10kg+20kg+40kg
+80kg-150kg
direct control
$3,650.00 $20,300.00
DYNAMOMETER WITH
CONTROLLER
D.C. Cradled Dynamometer
with Controller, 15hp
absorption, llhp motoring
l,000rpm-100km/h)
$24,200.00
H
$25,300.00
D.C. Cradled Dynamometer
with Controller, 17hp
absorption, 13hp motoring
l.OOOrpm (lOOkm/h)
$27,700.00
D.C. Cradled Dynamometer
with Controller 20hp
absorption, 15hp motoring
l.OOOrpm (lOOkm/h)
$26,850.00
15hp absorption
l.OOOrpm model TW-55
cooling water 6L/min
$15,000.00
7-12
-------
Table 7-3. PRICE AND SPECIFICATION FOR DURABILITY DYNAMOMETER
ROLLER
(Bl) D.C. Chassis Dynamometer Steel made,
with Flywheels 530.5mm
Total Price: $50,050.00 $3,650.00
(B2) D.C. Chassis Dynamometer
with two Flywheel and "
70kg Electrical Intertia
Simulation
Total Price: $40,500.00 $3,650.00
(B3) D.C. Chassis Dynamometer
with one Flywheel and
150kg Electrical Inertia
Simulation
Total Price: $40,150.00 $3,650.00
(B4) D.C. Chassis Dynamometer
with 300kg Electrical "
Inertia Simulation
Total Price: $42,830.00 $3,650.00
(B5) Eddy Current Dynamometer
with Flywheel only "
Total Price: $38,950.00 $3,650.00
FLYWHEEL
10kg+20kg+40kg
+80kg+150kg
direct control
$20,300.00
80kg+150kg
direct control
$10,000.00
150kg
remote Ccontrol
$4,600.00
180kg including
roller and dynamometer
i nerti a
$770.00
10kg+20kg+40kg
+80kg+160kg
direct control
$20,300.00
DYNAMOMETER WITH
CONTROLLER
D.C. Cradled Dynamometer
with Controller, 25hp
absorption, 19 motoring
l,200rpm (120km/h)
$26,100.00
II
$26,850.00
D.C. Cradled Dynamometer
with Controller, 45hp
absorption, 33hp motoring
l,200rpm (120kg/h)
$31,900.00
D.C. Cradled Dynamometer
with Controller, 50hp
absorption, 37hp motoring
750/1, 200rpm (75/120km/h)
$38,400.00
Eddy Current Cradled
Dynamometer with Con-
troller, 25hp absorption
l,200rpm Model TW-55
cooling water 9L/min.
$15,000.00
7-13
-------
There are several advantages to utilizing flywheels
which improve the performance of this dynamometer, as compared
to a dynamometer using electrical inertia simulation. These
i nclude:
t Operation of the DC system only in the regen-
erative mode resulting in faster response
times.
• Exact inertia simulation, regardless of
vehicle speed or mode of operation.
• Smaller power absorption capacity.
• Improved performance repeatability.
• Motoring capability.
When using flywheels to simulate inertia, it may
be necessary at times to motor the dynamometer to compensate
for windage and friction losses in the flywheel system.
Clayton has estimated that load loses as great as 3 kw may
occur, depending on the number of flywheels engaged. This
loss may be comparable to the motorcycle loads which occur
at low speeds. The compensation method would require the
determination of the system-loss characteristics as a func-
tion of rpm for each of the 60 flywheel combinations. A
driving torque could then be applied by the motor according
to a predetermined program based upon instantaneous rpm and
the position of the inertial selection switch. This system
would not be complex and would not require any modification
to the motor and only minimal controller modifications.
As a result of the addition of flywheels, additional
fail-safe provisions would be added. These include enclosing
the flywheel assembly and installing an interlock to prevent
7-14
-------
engagement or disengagement of flywheels during dynamometer
operation or in the event of power or air interruption.
There would also be an overspeed warning system.
Operation
The operation of the two DC systems is almost
identical. Personnel training efforts would be comparable,
and data acquisition requirements and set-up procedures
would be identical. Calibration procedures would be
simplified.
A shop air supply would be required to operate the
flywheel engagement system. This requires a compressed air
supply having a minimum pressure of 413.7 kPA (60 psig).
Power requirements are comparable to the other J3C
system, but AC current requirements might be slightly reduced.
Provisions must be made to return the regenerated power to
the line.
Calibration procedures, such as coastdown-tests,
are well-defined for dynamometers equipped with flywheels
and can easily be adapted for use with this sytem. Since
the dynamometer possesses motoring capabilities, the coast-
down procedure can be easily automated for little additional
cost.
Spatial requirements are maximized with this
system. This occurs because the flywheels are significantly
larger than the roller and DC motor.
Maintenance-
The complexity of the controller is reduced by
utilizing flywheels for inertia simulation, and this should
result in a more reliable system. However, with the introduc'
tion of the flywheels, additional maintenance requirements
are necessary. These include periodic lubrication and the
7-15
-------
inspection and possible replacement of bearings. The level
of maintenance is comparable to the other DC configuration.
Costs
The purchase price of a DC system with flywheels
is significantly higher than the other DC system. This
occurs even though the capacity of the DC systems using
flywheels is significantly lower as shown in Tables 7-2
and 7-3. The added costs of the flywheels more than offset
the reduced costs of the smaller regenerative motor. As
with the other DC system, there is a substantial cost differ-
ential between a certification dynamometer and a durability
dynamometer.
Installation and maintenance costs of the two DC
systems are comparable. However, the operational costs of
the DC-flywheel system will be slightly lower than the costs
of the DC-electrical inertia system because of the lower
power requirements of the former systems.
Components of this system are standard products,
but are not necessarily stock items. Estimated component
delivery times are 12 to 26 weeks after orders are placed.
7.5.3 Eddy Current PAU with Flywheels
Performance
The eddy current system consists of an eddy cur-
rent brake, flywheel assembly, and controller. The controller
and flywheel assembly function in a similar manner to the
components used in the DC-flywheel systems. The controller
will provide a DC-excitation signal for adjusting the clutch
field coil on the eddy current brake.
The response times of the eddy current unit will
be slightly slower than its DC counterpart. The computation
7-16
-------
and picking times within the controller will be on the order
of 100 milliseconds, and the response time of the PAU to
changes in the excitation signal is estimated to the 500 to
1,000 milliseconds. However, quantitative information on
the response-time characteristics of eddy current brakes
could not be obtained from manufacturers.
The power absorption characteristics of an eddy
currrent absorber at low rotational speeds, less than 100
rpm, is critical in this application. Available data is
skimpy. The Lear Siegler eddy current brakes, examined in
Section 2, are not effective power absorbers at these condi-
tions. On the other hand, data published by the Dynamatic
Division of Eaton Corporation in their publication, "Dyna-
mometer Handbook of Basic Theory and Applications," indicates
that eddy current brakes can be effective absorbers at low
rotational speeds. A typical set of torque curves is shown
in Figure 7-1.
The eddy current flywheel system is equipped with
comparable fail-safe provisions as the other configurations.
Should a water-cooled brake be used, as is recommended
below, additional interlocks willl be required to protect
against water/temperature failures.
The repeatability of the system's performance is
imperative. The use of flywheels assures the exact simula-
tion of inertia. In order to ensure the repeatable performance
of the eddy current unit, it should be water-cooled. This
eliminates performance fluctuations caused by motor and
rotor temperature changes. The repeatability of the control-
ler will be less than 0.5 percent.
Operation
The major difference between the operation of this
system and that of the other two systems is the requirement
for water cooling. A typical cooling configuration is shown
7-17
-------
>
I
I—'
CO
-
0
DC
0
IL
C
LU
o
ff
10 20 30 40 50 60 70 80 90
PERCENT OF FULL SPEED (10.000RPM MAX)
1 0 0
FIGURE 7-1.
TYPICAL TORQUE CURVE OF STANDARD DYNAMATIC DYNAMOMETER
(EDDY-CURRENT TYPE)
-------
in Figure 7-2. A closed-loop, water-to-water heat exchanger
is recommended in order to minimize thermal loads in the
test eel 1.
Personnel training and data acquisition requirements
would be similar to those required in the other configurations.
Additional steps would have to be added to the set-up proce-
dures to ensure that the cooling water system had been
started. The eddy current brake is only a power absorber;
it is incapable of motoring. Should a motoring mode be
required so as to perform automated coast-down tests or
windage compensation, an eddy brake can be coupled to an
eddy current motor. This system is available from both
Eaton and Louis Allis. The cost/performance of the eddy
current motor/absorber versus a regenerative DC drive have
been examined in Section 6.
Spatial requirements would be comparable to the
DC-flywheel system. While the eddy current brake is smaller
than the DC unit, added space will be required for the heat
exchanger and related plumbing.
Maintenance
The eddy current system is the least complicated
of the three alternatives. The brake unit has less moving
parts than a comparable DC unit. There are no brushes,
commutaters, etc., which require replacement. There are
bearings in the unit which require periodic lubrication.
The major maintenance item in this design is the
cooling system. The cooling system is required to dissipate
heat from the eddy current brake. A closed-loop system with
a water-to-water heat exchanger is recommended in order to
minimize water usage, which is approximately 644 1iters/min/kw
(12.7 gal/min/hp), and reduce the heat loss to the laboratory
environment. Since the water and its natural contaminants
can corrode the internal parts of the eddy current unit,
certain precautions should be taken. These include:
7-19
-------
WATER TO WATER HEAT EXCHANGERS
MAGNETIC DRIVE
RAW WATER INPUT,
CUSTOMER HAND OR
REGULATOR VALVE
RAW WATER
DISCHARGE
COOLANT
CONTROLS.
OVER-TEMPERATURE
SWITCH-
AIR VENT AND
WATER FILL
MOTOR AND PUMP
PIPING SYSTEM FOR A MAGNETIC DRIVE WITH HEAT EXCHANGER
FIGURE 7-2. EDDY CURRENT BRAKE COOLING SYSTEM
7-20
-------
• Coating the inner surfaces of the brake with
an epoxy coating.
• Installation of strainers to remove parti-
culate matter.
t Use of 50 percent water and antifreeze solu-
tion to minimize corrosion.
With the proper maintenance, the eddy current
power absorber has a useful life in excess of 20 years.
Costs
As shown in Tables 7-2 and 7-3, the eddy current
system is the least expensive of the three alternatives.
There is no cost penalty incurred if the system is designed
to meet the objectives of the more demanding durability
cycle. A severe cost penalty is introduced if a motoring
capability is required. Then, the costs are compatible to a
DC system with flywheels.
Installation and operating costs are comparable to
the other alternatives examined. Also, delivery times are
consistent with the other alternatives.
7.6 CONCLUSIONS AND RECOMMENDATIONS
The results of this study are tabulated in Table
7-1. As shown there, the eddy current flywheel system is
ranked first. The second-ranked alternative is a DC system
employing flywheels for inertia simulation. The least
attractive method is a DC system utilizing both electrical
and mechanical techniques for inertia simulation.
7-21
-------
With respect to performance, the DC system with
flywheels and the eddy current system with flywheels are
comparable. The latter system has been recommended because
of its lower price, reduced maintenance requirements, and
higher reliability. Should motoring capability be required,
these advantages would be virtually eliminated. There is no
technical information in this study which justifies the
exclusion of a DC-flywheel system.
The use of electrical techniques for simulating
inertia would be acceptable for use in a mileage accumulation
dynamometer. However, it should not be used in the certifica
tion dynamometer because of its lengthened response time
characteristics, potential inaccuracies which will vary with
the motorcycle mass, and complex calibration procedures.
Since flywheels are a primary inertia standard, there is no
justification for simulating inertia by electrical means.
The discussions which have been presented in this
report have been qualitative in nature. Throughout this
program, dynamometer manufactureres have been contacted in
an effort to obtain quantitative information on which our
recommendations could be based. The data have not always
been forthcoming. It is probable that, for the most part,
the data have never been generated. If the information is
known, it is considered proprietary by the manufacturers.
However, for the purpose of recommending a design approach,
it is believed that sufficient information was obtained to
develop a firm qualitative analysis.
7-22
-------
Appendix A
MOTORCYCLE TEST SUMMARY
-------
jWJ JU3333)
6.5 J
;\'t.*l?2,
FLK. •& CF ZERO SKIFT.
J
it!
\x /Zfwes
j ;
'~ \ _ -
p;
•' !
|_1
_ 4 ...
J
, .„
, |
. •
— ^
I
— - I
FE5T" i- EllTlFil'fiTlCU i
- AWKf r |WiiFaiaS|RaM?.fl. i 'uswc| «t /^, .1^, .~ ' * , " • l^c
, CFJ . ! i f._wj ;Tnw) ; CRI« ^ ' fFF»;) !.,>/-:>_:. *.- v ,>-MI
, 1 .... ,. t r )• ]~/700'- 3.<£f\'/,Z4 7f&
i ' i '' ' : | "" j I' //2? :/5./5 j 3-25" ^5"
':"" " i" ; ! " " '' \ i 7'? . — \.~ \ 28^
i • I tfSOo \ ~ ' iT*J? /
""' "•"" i " " ! : ! f 45? i — : - /tf3
;- j— j j- ; | ;
, ' . ! i , i • f. A £Z\ _ , _ ' fff> a
— **" ' fJpffSA^Q- £ \ &•&•£ \ ££•£"& ' *i i I- ^^~-f \ . • D j
- : J 1 1 i. 1570 i — : ~" 62e
r '.'']'' • " • l! j&4d\ 3.35 ;./-^5 6(g^
— • ; H— -f .. 'ii //7?^. w^r -3.24 ! <57
- !-- r--t f -•: !! . ! S
; I ! ! '• 365 ~~ ' /£&
..-,.,.. .-1 - ) - 1 || '; '
! , - • - j - - : r i : !
i , ! | • !j..,.i
• • e ' 1 ! 1 . . . |! j^£ — — : j£Q
^-•—!***-r z j-- - • -]•• -; |>73:r -".: - "|«/
!- ' 'T "" "; "" "" " 7 : li-/<5/5- c-35 : '-9-T-i- 6-53
r J- ~! * T i \J.I.7C> \I5-V° I 3.34 483
i ' (i i . :
- ";/aa*»' 1 • *._/6'30.3.&9.43'5.
• ' " ": ' "j : 38) ,. ..-. .— . /5<£.
— • • •— ' ' ' -j- • . ' t£>C&\ - , 4£>6 .
t- :•;•-•--: ^i - i.,..."...
c- • \~ ', ~~ '"' g 6f &S6 20 \ \.-3S l\ " .. ~~ l^> '-
— •-- —i — - . - • r- -; ----'-— -.--.-•- - - if C~ £T^ f 1 !? 7 /)fa ' f f S\
* - i i . / c>*> ^y ' (0*23 ^ ' *s^ ' [p fou
"j i ' . 'I J. i.
A-C Ttr:,.{
r * f r j
16. CO
/?.40
Z4.BC
14. OO
14 AO
JS.60
J5.2C
I3.B6
{4.4.0
I SAO
13.12
22J2
S2.2O
12.80
fZ.SO
14.68
30.20
. SC.48
...23.2C
*%lo
: "'
• I2.<30
. 36.00
• - —
. _ - .L '
A!C rj/.VV A TTf.J 6K.-;U U«
MC-FflU MC-TAJgfe
inn itf', vj HS
j?'A2—*-Sj:/?—'?--:?£ -A ;
1-53 f J7.S5 ' J'l? A '
3.35 : £'•}£ 2.14 F )
4.57 ^.-f5" 1-32 C ;
0.77 ; — . ~ .- i
/.2i) - ~ -
A/9 , - - -
A61? - —
Z./4 ! — ~ -
O.^fi: - — .-.
450' - - ~
A57 ' - \ - -
"2. .76 . /ft 77 /.73 S
-3-£^ 25.^7 Z.CS' S
o.3s. - . — ;-.
?.'/
-------
MO^ELVCLfl "TEST SL//y-^v
HbfrODRTfi fSf?REC:ED FDR Vi t)F ZCPJ SHIFT.
2) LlKL LETT£« DEMTEAFPRaLSAME MC LOUD INDEX.
5! t .f.. ..r . . . • - L Kpaft. j,. crOi,
}=* /'Fj rwj nun : m>ia}
l?l/<5>M"i3' ££ \f/0"WX> Z R-15 t7.£? I ZC '
I i ' • ; ; r
..:.._-!. i ; . !
J_"j. "7.^.1 .._.j j .. ;. ...
i 6 \iSMC75
•• 1
r- •• - -
.7 \i5fUfa
•
: 1
""
' '
!
' : i
. - - . A .. ' . i
,_._ ' .. V. i" "" ,
4- ... |
* • !
-.. ' i
1 : - » — -i
1 -
J
.70. .jM«WM^../_. . j_/?.7J
]
J j ' !
JEW*
!
^^ /jffjftfj) Jf t ' 20* 0&
- ;! ._ i_ :..
1 . ' '
i * - - -
Ll^riHf&S
-«-' -*«**• - -ze°°
j I ! .
t • •
/7. £9 I /4
i '
i
i \
\
: i. 1
i
i
! |
i
1 i
-- \.»..
\ J I ' !
j 1 '. .._.._ - .._ ..t
1 '
1 , : ;
i
1 ! : .1.. 1 i
COMS7Vi/i7.3 : W.U'i'/i H'.-if CL ATM - CCl/L'WT Jf /.3333)
HflFtEVD. lt'HF£i.S>"f.;- fCi)!.'N7Jr I.,'7£S J
^
.
•
i
;
! •
<=»
rflii ijj,"^
794
/£>£/
. 77g
77 633 ! 29.30
2f.5» i 3.^-5 -f^7 ! ^?-2<5
- j - IS! \ °>-20
- i - 47; : 10-10
i !
- ; - /?? ; 9.28
- i - { -m i 9.00
//.60 , 2.29 • £2g ;27.go
^-?.2f ; 3.73 4^3 ;4g.^
- - I * t
~~ - i - /^e ! 9J2
- . ! ' 45^9 ; 7.48
je.OS i ^-40 £20 .; 2J.52
23.9(5. 2-9^ 6/5" 29-92
4./.3JT , J.S9 47J 47.20
- 1 - .. 387. .6-92.
..... i .... ..:
..- ' jr. 59/. 1.9-72
- .'. - 74«?.. #.V2
20-.4.5 .. .2.^9. 635" .1 20.O8
/^•6^_O-93 4// 1/7.O6
0.99 _ —.!..— .-
4.46 zsfto- O.Sd> c.
3.5*) • 19.00 • /.40 ft
4.64 27-7O A&5" C
0-26 i - ; -
0-77 j - ' — -
3-3Z , I&-20 . l.07> B
4.54 I 24. JS [ 0-ei £
3.31 ' I7.5O . l./G B
4-3& Z2.10 ', 0.59 ,C
.0.26' '— \ — '-
\ '
3.01 ' &.4?' O.&t 3
4-\°) . 5.5O 0-29 r.
3.50 . &.OZ o.ss y
4.25 5.85 0-36 c_
. — . . .
0.65"" - ' " -
.1-14.1 - ... - '
3.40 \. 7.£3. 0.81 B
/. 34 i 6.4& 0.41 B
_4:25| 5.?e.o.3l.jc
-------
LEi TEST SLJ.W/!^
FA.J
ZTMAZJf PHMJ
'.V6*i.32
El roe Vt OF
LErrees
TEST i:.£;;riFft:*T/fiw
MC LC/ID INDEX.
1 ;
- a.tru -
l!1*15
(s i tames
i j
: 1 .
,_..
L.;***^
^ i^
t
r ?
/s aofoezs
,
' i& ZJ/VH7S
' 1
1
'.".mi
. ; — -
^
!>*HA7Bl5rlwlZtritiFAn.JFaiKis} K/.L1WA.
\ \
f'FJ ' 1 ' f'«l
tSUUSFltt
(iwi
7^ JPKSJkftai f ' Z0.6O —
J
. . 1. ; ..:
i ; \ •
Gff ^ftofiM^ / j 2f>jffo__
- '. '• \ -
t 1
7/ jMWJ / !.»•«»
t <
r ;
71 I/6B /=2y / ' 12- 73
. .I... ...._.; . .
""i" " i ;
• - ! ' 1 " "
i — . -» j
\- *- •
.._
_"
! taiL j . t
i 26- ' :l 7^>4 ' 34.45"' 4.^2
1 .
,.. .. .(
i
!
•
|
(
- **-'-!
i
*
•
• i
_
_2-tf
.. .... .
1
_. _ .._ . -,
\~I901 3-2S\ Q.12
939 \ IIB'IO \ Z2A&
— * * ' f • j *• "
toOS \IQ7.2S' 12.35
r- ;
^ &y=) i 32.30; 4.30
\ 42J ; /Z.'Sf' /.
^ r f*s | •- , ^
| &02. \ 41.30 j 4-73
i 3^8 | /3.05" j 0-86
i j
-.699 =37. 7f i 4.23
f 17$ ?-.20\ 0.07
llCC? \i07.C.f\ 2.0.6,8
1 £02 37.4S '!_ 4.2e
^ 326 i IO-6-0 \ Q-(>t>
• ,~
'•422 13-45. 1.08
; 177 . 2-g£>. 0.<39
i 1001 ' 1 30.95. 25. //
r<£C9 [ 4&.70\ 5.-J2
^.•439 i 22.90; /.9/
MC 3f^£P JAV: TTKrf *WC PJ.'.'flE A
• ' • w
JK7.1J f*FH : WD i
S821S3.7Z } f.?2l
372^\25.eS_
/S& \13.I8
823 \I63.70
507 1/46-97
585 53.21
360 ; 27.87
/47 j 13.18
SOS\ 66.H
]
368 \ 20.63
841 \/6l.l6
£02 _\ &J-32L
273 ' 24. g3
579 1 57.^2.
348 2C.AS
'.840. I67.O3
_5G6 ; 65- €8
3£3 } 37.50
.
i_ /(97^ ^ .^2. 05;. 4.49 .1 5.73 . 52. 96
'' 77^? i 10-30'.. /••£"/ i -?/./ „ 3M2.
! ^95";. 3.^5, 0-23 j 2/4... /7.4g
1/59^ 82.65: 25. 'Oj 058.^7^9
I..9^J \-2?.4G... 5-37 | ^"/J 1 £.3,60
i 1 •
! _...].. .. . ;...._
../•£3._
O.39 ,
2.S.8I . -
26*^3 ' -
6.36 1 <
2.01 \ >
0.36, ;
25. 5/ ; .
5.56; .
ATf
C3.36
6.35.
'
~2-4&.
2.7.30
.2-381.
I ~ [
L.1.1
-a m « '
MC-FTiil
V'
0.27
3-35
3. 0.2g
46.60, 3.99
24.04* /.63
15.56 0-73
£ '
£ ,
23.97 _ 1.57
14-23. 0.63
27/7 . 2.?7
13.00 . 0-tf
to. 12. O-27
.
, 0-29 P
T>
V
P
£
. .
2/./Z ; 0-95 J?
y4.4J O.4S I?
84.44 2.20 £•
34,20..
_-..
-------
FA.J Cflti
NOTE-'lWfiTA COZKECTED FOR '/a OF ZCBS SHIFT.
i) UKE LETTER DENOTE APPKOX. Sf.lf.E MC LOflD INDEX.
TL2r ILI
r.ires j
fci, 9.65D = CMfnl=
12.7- D - (t-'.rn'i
' ?f.C,'.n - WfJC l
73
DATE
[A? ZIHfitTS
1 ;
n- —
i
fc"~* — -• ••
— ~. —
"
•
" . .'
.
J»J8W
,
3_24taa
i
„ i
£5 4#.&fW\ 2 /2.7S
r ^ !
71 !*A«*' / 1 ^.
'i
A38 i
37 .
O.I8 !,2/3.;
29.4 i
37.36 i
ZZ.&f
tS.Z*
. /£ .^7
'' 2&.7I
i.ez-
>. 3°> B
Q.~6 &
&.-B s.
\,3°) 6
6.20 6
0.41 O
•-—(--.1 --
-------
FAJ KOU.
e.'.J
NOTE: IJIftrA CORRECTED FDR V4 OF Z£KQ SHIFT.
n LIKE. LETTEK KNOTE APPROX. SAME W£ LOAD IHDE.X .
>JijE*T£ t*e.TEftyp(JMfiafK-fAc.af"a3U5jzatLp«fl. |s5us/tfri77«'r/£si "wiiy
ia: ; i i t TG t-t KOU.
SKI ffj i ! ! rwi i nun i «%ico
\ ...',. 'IJ**6 \ '-GO j 0.35
'. {.. .,_ 1.377 \ O.50\0.04
."('_!'. ?' : i
: - 1 zo ' i t>8
__ ! ._.. 145
— ; ji '*<
!....{. t i &6
_ 1. . 1 j 44
I I /s,
'• • • 1- -
-"--- :; f-
" " , . .;
'_" ' -U L_
. J._._ ij |T.,
ir •••
J i ii
-- --., - -1 ^
C- -
fL_
••*• •• •"; F
. _. _ '... - -
j - 1-
L-
jt j
;• j
— __~- . . , . — U j — — • '•
1 ! ^
: . i
:>.~~t.~~ """i -—': -~ \. •-;:;.., ' _ - i
1 \ 2O.60\ 2.&7
7 : IQ.3O \ O.^O
3 "| 2.5$ \ 0- OB
6 ,30,00 \ 3.80
5 14.00 ' 1.19
6\ 2.40 \ 0.07
t
t
f
•" 1" '
.1. .._ .. i. . .
1 ;
...J..... .;.. _
i... ..i ........
! 1- 1
' '- rwpwjc ""in SHEET 5.0F5
MC yns I/HL 7iE;f 'we PJ/.VT a asi/t & ttam L>*
| : *)C • F»J Mf-fiW ^
G>3\ 1 35- 9£ -^.32 ' 28-56, Z.IO 0 \ I5^S5- O.47 6
\. 3.43\ 2398,' 2.04 S
1-53 • 17.86. Lie .S
0.45~\ 15-38 \ 0-41 3
3.47. 1 7.92 i O.BO j?
/•39 ' 6.^6 ; 0.4<) J
O.24 • 5.46 < O.)6 £
4.30 ! £.04 : 0-50 6
I.5B\ 576 . 0.3*5 6
O-22 '. S-M i C.I5 .G
; i
i ; ' i
,. . . -- , - --- -
...-,. . .
i - • i
, . -- . ,
- -
• * \
j
" i " " '
! i
t ;
i - -
...| — — —
_.;._. ... :
TT:T< . T:."
. . _
,
! ...
-i •- - •- ;
— ; . - L
i !
._ . . . _j .L ..... ...l
-------
TECHNICAL REPORT DATA
(Please readlitsiniciions on the reverse before completing)
i. REPORT NO.
EPA-460/3-76-004-A
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Development of Specifications for a Motorcycle
Dynamometer and Motorcycle Cooling System - Vol.
Design Study
5. REPORT DATE
Feb. 1976 (date of issue)
6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)
8. PERFORMING ORGANIZATION REPORT NO.
Robert J. Herling
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Olson Laboratories, Inc.
421 East Cerritos Avenue
Anaheim, California 92805
2 AB 130
11. CONTRACT/GRANT NO.
68-03-2141
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air and Waste Management
Office of Mobile Source Air Pollution Control
Emissions Control Technology Division
13. TYPE OF REPORT AND PERIOD COVERED
Final 9/74-2/76
14. SPONSORING AGENCY CODE
is. SUPPLEMENTARY NOTES Ann Arbor, Michigan 48uJ
------- |