EPA-AA-TEB-81-21
Evaluation of a Dresserator System Test Vehicle
June, 1981
by
Thomas J. Penninga
Test and Evaluation Branch
Emission Control Technology Division
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
-------
Abstract
A test vehicle supplied by Dresser Industries was tested at the EPA Motor
Vehicle Emission Laboratory to determine the feasibility of the Dresser
Sonic Flow Carburetor system as applied to an 3-way catalyst system. The
testing conducted included the standard Federal Testing Procedure,
Highway Fuel Economy Testing, testing at 20°, 40°, 60°, and 70°F and
sulfate testing. The test vehicle achieved emission levels below the
1981 and subsequent model year standards of .41 gm/mile hydrocarbon, 3.4
gm/mile carbon monoxide and 1.0 gm/mile NOx. The vehicle suffered
starting problems at lower temperatures but had no driveability problems
when warmed up. An extended idle period at the beginning of the cold
start test procedure was also used at lower temperature. This modified
FTP procedure improved driveability and lowered vehicle emissions
somewhat.
Background
The EPA is interested in analyzing current automotive technology to
determine the effects of such technology on emissions and fuel economy.
The Dresserator System has been claimed to markedly reduce automobile
emissions. Testing by several laboratories including Dresser, General
Motors, and California Air Resources Board have substantiated these
claims. Therefore, EPA requested that the prototype vehicle from Dresser
Industries be made available for EPA testing at the EPA Motor Vehicle
Emission Laboratory.
Test Procedure
A test plan was submitted by the Characterization Technology Assessment
Branch for approval. The test plan calls for Clayton split-roll
dynamometer testing us.ing Federal Test Procedure (FTP), and the Highway
Fuel Economy Test (HFET), and the Congested Freeway Driving Schedule
(CFDS). The Clayton testing, was followed by controlled Environmental
Test Chamber (CETC) testing at 20°, 40°, 60°, and 75°F and using both
standard FTP and HFET sequences and an extended idle FTP sequence. The
CETC uses a single roll Labeco dynamometer. The extended idle test
involved was a standard FTP with 45 seconds of additional idle prior to
the start of bag 1.
Device Description
A complete description of the Dresser Carburetor and Dresserator System
including schematics and supporting test data was supplied by Dresser
Industries. This information is also included in Attachment A. The
Dresser information regarding a sonic EGR value was deleted since the
vehicle tested by EPA did not have a sonic EGR valve.
Test Results
The test results are presented below:
-------
Test
Number
80-8112
80-8149
Date
3-10-81
3-12-81
-2-
Table One
Standard Federal Test Procedures on Clayton Dynamometer
HC CO NOx FTP
(gram/mile) (gram/mile) (gram/mile) miles per gallon
.2394
.2180
1.889
1.795
.3136
.3078
13.482
13.265
80-8113
80-8150
Table Two
Standard Highway Fuel Economy Test on Clayton Dynamometer
3-10-80
3-10-80
.0169
.0180
.327
.440
.0919
.1109
18.302
18.184
These results generally agree with the claims made about the Dresserator
System and correlate with other data generated on this test vehicles.
The FTP results show emission levels below the 1981 and subsequent model
year standards for HC, CO, and NOx are quite easily achieved. It is
important to note that this test vehicle is a large (4000#) vehicle with
a large (350 C1D) V-8 engine. It is our judgement that smaller vehicle
engines combinations could utilize the same system and achieve equivalent
emission results.
The vehicle was then transferred to the Controlled Environmental Test
Chamber (CETC) for testing. It must be noted that the Dresserator System
was not optimized for low temperature testing. Therefore, cold start
testing problems were not unexpected.
The Dresserator vehicle with a standard FTP starting idle period would
not run properly at 20°F. After two attempts at 20°F with over 6 stalls
in the first 20 seconds, the decision to abort 20°F standard FTPs was
made. The testing plan was modified to add 45 seconds to the initial
idle period to hopefully prevent the stalling problem. The results of
the standard FTP, extended idle FTP, and HFET tests are given below:
Table 3
Standard FTP in CETC
Test
Number
80-8195
80-8189
80-8191
80-8187
80-8205
80-8181
80-8183
Hydrocarbon CO
Date (gram/mile) (gram/mile)
3-24-81
3-19-81
3-23-81
3-19-81
3-30-81
3-17-81
3-17-81
.5624
.2836
.3005
.2584
.2548
.2570
.2640
4.092
2.826
3.258
2.001
2.875
1.380
1.221
Miles
NOx per Test
(gram/mile) gallon Temperature
.7588
.5664
.5296
.4302
.4330
.3582
.3228
13.351
13.492
13.214
13.376
14.002
13.643
13.841
28.0°F
40.0°F
40.0°F
60.0°F
60.0°F
75.0°F
75.0°F
Number
of
Stalls
2
0
1
0
0
0
0
-------
-3-
Table 4
Extended Idle FTP in CETC
Test
Number
80-8193
80-8199
80-8200
80-820.2
80-8445
80-8464
80-8463
Hydrocarbon CO
Date (gram/mile) (gram/mile)
3-24-81
3-25-81
3-25-81
3-26-81
3-31-81
4-1-81
3-31-81
1.6927
.8807
.2787
.4970
.2033
.2683
.2570
8.098
6.361
1.770
3.490
2.345
2.836
1.678
NOx Miles per Test
(gram/mile) gallon Temperature Stalls
.8996
.7979
.5674
.4825
.4195
.3847
.3467
12.722
12.855
13.422
13.156
14.206
13.638
13.447
20.0
20.0
40.0
40.0
60.0
60.0
75.0
10
7
0
4
0
0
1
Table 5
Highway Fuel Economy Test in CETC
Test
Number of
Number Date
Hydrocarbon CO NOx Miles per Test
(gram/mile) (gram/mile) (gram/mile) gallon Temperature Stalls
80-8194
80-8196
80-8192
80-8201*
80-8190
80-8186
80-8188
80-8203
80-8454
80-8182
80-8184
3-24-81
3-25-81
3-25-81
3-25-81
3-19-81
3-19-81
3-18-81
3-30-81
3-31-81
3-17-81
3-18-81
.0262
.0281
.0297
.0475
.0316
.0228
.6255
.0181
.0157
.0231
.0188
.534
.478
.659
1.180
.796
.489
.360
.393
.425
.353
.400
.1870
.1569
.1924
.1558
.1588
.1684
.1230
.1516
.1235
.1026
.1079
18.480
18.753
18.506
19.028
18.735
18.636
18.720
18.881
18.721
18.410
18.682
20.0
20.0
40.0
40.0
42.0
60.0
60.0
60.0
60.0
73.0
75.0
0
0
0
1
0
0
0
0
0
0
0
*Vehicle lost power and stalled at 200 sees into sample bag. No reason
was found.
As can be seen from the data, the low temperature cold-start stalling
problems caused high variability in HC and CO emissions. This was not
unexpected. The data shows a direct correlation between the number of
stalls and the HC and CO readings. The temperature data indicate that
HC, CO, and NOx rise as the temperature is reduced. This trend is
consistent with other ambient temperature studies run in the CETC.
It is interesting to note that the extended idle FTP data is not
significantly different than the standard FTP data. The extended idle
did not noticeably reduce emissions or improve fuel economy. This
comparison is probably masked by the stalling problem so numerical
comparison would not be realistic.
The sulfate samples taken during the CFDS tests have not yet been
analyzed. The necessary equipment to make the analysis is under repair.
-------
-4-
When repairs are completed, the sulfate data will be tabulated. Copies
of the sulfate data will be available from the TEB secretary.
Conclusions
The Dresserator System performed according to the claims made about it on
the test vehicle supplied. The low temperature driveability problems
must be addressed before such a system is put into production.
-------
Attachment A
INTRODUCTION
The Dresserator Inductor system on a 1977 Chevrolet
Nova has now met the statutory standards of 0.41 HC, 3.4 CO,
and 0.4 NOx, while simultaneously providing an economy increase
of about 5% over the base car. The addition of another Dresser
development added an additional 5% economy increase at these low
emission levels for a total of 10% over the base car. The
Dresserator Inductor system is simple and commercially pro-
ducible; it has a cost advantage over other systems designed
to operate at low NOx levels. The present system utilizes a single
three-way catalyst with no air pump or oxidation cleanup catalyst.
It is anticipated that the system can operate at the
1 gpm NOx level of the 1981 standards without the necessity of
EGR, providing a further cost benefit. A vastly simplified
version of the overall system has considerable economic advan-
tage in operation at still higher NOx levels, as encountered in
Europe. Under these conditions, the system is operated lean
with a considerable (10-20%) increase in economy over the base
car.
The heart of the Dresserator Inductor system is the
Inductor itself, a variable critical-flow venturi. The design
of the venturi enables it to maintain sonic velocity at its
throat over most of the driving range of the car .-^Ond-e-j?—these
conditions, it can control mass flow and is an excellent atomizer.
The mass flow control capability of the Inductor is used by
Dresser in a variety of valve designs, one of which is a sonic
EGR control valve which enables one to have a simple programmable
EGR system. The atomization feature is utilized when the sonic
principle is used as a carburetor. Here, fuel is added above
the throat and excellent atomization is achieved as the fuel
passes through the sonic throat. Since all of the air and all
of the fuel pass through this throat, mixing is inherently very
good. Many atomizers do well under steady-state conditions but
-------
it is this ability to atomize and mix well under all "conditions
that separates the Dresserator Inductor from the pack and permits
attainment of these unique results.
Although not used in the current Dresserator Inductor
system because of time constraints on our development' program,
the mass flow control capability can be utilized to"further
simplify the electronic control system since throat opening
is a direct measure of mass flow to the engine. In addition,
the mass-flow control of the sonic EGR valve can be incorporated
to provide a simple, yet highly sophisticated system of engine
control.
This report describes the Inductor system used and the
results obtained with the 1977 Chevrolet Nova equipped with a
350 in engine. The general principles of the Dresserator
Inductor are described and comparisons made with competitive
systems. In addition, the advanced Dresserator system is also
discussed.
-------
TECHNICAL DEVELOPMENT AND RESULTS
The Concept
The heart of the Dresserator Inductor system is the
Inductor itself. It is a variable critical-flow venturi. By
critical flow, one means that the velocity at the throat is sonic
at which point the mass flow cannot be exceeded provided the up-
stream conditions of the flow remain the same. Variations
which occur downstream, or in this case in the intake manifold,
have no affect on flow through the throat, provided sonic
velocity is maintained. The ability to maintain sonic velocity
at the throat over a wide range of manifold vacuums is the unique
feature of the Dresserator Inductor. This ability is accomplished
through the utilization of a diffuser below the throat which con-
verts the high velocity flow energy to pressure. By proper
design, one can achieve high energy recoveries in the diffuser
which allows operation from high vacuum to the range of 3 to 4
inches of manifold vacuum, while maintaining sonic flow at the
throat over the full flow demand of the engine.
Figure 2 shows that a critical-flow venturi consists of
an entrance zone which is subsonic, the throat which is sonic ,
and, depending on manifold vacuum, a supersonic zone,followed by
the subsonic zone. At high manifold vacuum, this supersonic zone
can be quite significant, extending for an inch or more below the
throat. This supersonic zone usually ends in a shock after which
the flow is diffused for the remainder of the length of the diffuser.
Under these conditions, energy recovery is not important since
the unit is operating as a throttle wiuth a significant pressure
differential. As the pressure differential decreases, the
* *
supersonic zone shortens until the shock reaches the throat at
which point, if the pressure differential continues to decrease,
the system' can no longer maintain sonic velocity and the throat
-------
r-\.
riff?' "of',M.;i,:' ~
•'£-*!•''• c'< ;?,f-:; '
sc
supersonc
-------
becomes subsonic and the system behaves as a modulatable, sub-
sonic venturi. In the case of the Dresserator Inductor, the
diffuser functions to recover energy from the high velocity flow
stream such that as the pressure differential decreases, the system
remains sonic even at very low pressure differentials.
Sonic velocity under ordinary conditions at- the throat
is around 1100 fps. This velocity can be utilized to create
a very large shear force on fuel particles as they pass through
the throat and cause the particles to break up into minute droplets
generally in the range of 10 micron average size. We have found
that it is important that the fuel be predistributed across the
qntire throat area in order to optimize the atomization. If this
is not done, the particle size produced is much larger than 10 micron
average and, indeed, can be quite non-uniform. Various predictive
equations on atomization confirm the approximate average particle
size that would be achieved with this kind of velocity and proper
fuel distribution. Work by Stanford Research Institute on some
of our earlier atomizers also confirmed this average particle
size. More recent work on the Dresserator principle by British
Leyland using very advanced particle size measuring techniques
indicates an average particle size of 6 microns for the Dresserator
Inductor. This average particle size is maintained constant down
to the range of one inch of manifold vacuum before the effect of
the lowered velocity of the throat is felt and th£^.ax±JLcJLe_size
begins to increase. Furthermore, British Leyland has confirmed
our evidence that less fuel is on the manifold walls with sonic
carburetion.
We have tested many atomisers and compared results with
those from the Inductor. In laboratory atomization tests, the
Inductor was always found to be superior. On the car at steady-
i
state conditions, it is often difficult to choose among good
atomizers. However, the ability of the Inductor to control
atomization over most of the manifold vacuum range and to mix
the fuel well with the air provides the significant differences
-------
10
that lead to the unique results obtained. When compared to
other carburetion systems, we have generally seen lower NOx
and when running in the lean-burn mode this can be as much
as 50% lower. We always see lower CO by as much as two-thirds,
Advantages of the Concept
Sonic carburetion has two important principles, which
differ from other types of carburetion. These are -
• mass flow control
• atomization control
When these two principles are properly applied, one then
achieves the benefits of sonic carburetion. These benefits are -
• excellent cylinder-to-cylinder distribution
• excellent cycle-by-cycle distribution
e lean cold start capability
• improved economy
« excellent air/fuel ratio control
Under most of the operating conditions encountered in
the CVS cycle, the Dresserator Inductor system on ^th,e_ 1977
Chevrolet led to a cylinder-to-cylinder distribution spread of
only a few tenths of an air-fuel ratio. Under conditions where
the system would become subsonic, this spread would increase to
the range of one air-fuel ratio and only under wide-open throttle
conditions at low RPM did it exceed one air-fuel ratio.
«
Cycle-by-cycle distribution is one of the features that
we feel has major significance when one is concerned with opera-
ting at stoichiometric with a three-way catalyst. Since the
-------
ll
sonic carburetor is an excellent pulsation dampener,'its pulse-
free charge provides excellent cycle-by-cycle air-fuel ratio
control. This we determine by following the break point of
the carbon monoxide emissions from individual cylinders as
the engine goes from a lean condition towards stoichiometric.
In general, we approach quite close to 15:1 A/F ratio before
the carbon monoxide will begin to rise. Whereas, with other
carburetion systems we have examined, this break point will
occur at a much higher air-fuel ratio, indicating that the
individual cylinder charging is quite variable. This can be
observed even under conditions where one is measuring excellent
cylinder-to-cylinder distribution.
We have long recognized the importance of our cold start
capability which is brought about by the excellent atomization
that is achieved. Since the average fuel particles are quite
small, more of the gasoline travels through the cylinders in the
air stream than with other carburetion systems which require exten-
sive choking for enrichment in order to achieve volatility of a
fraction of the fuel which then goes in a vaporized form into
the cylinders. This is a major reason for the control of carbon
monoxide emissions which has been recognized by others testing
sonic carburetion.
Under all conditions and with all cars wj^haye examined,
we have always observed an improvement of fuel economy over the
base car. In some cases, this has been as high as 30%; however,
in those extreme cases, we must readily admit that a part of
the reason was a very poor carburetion and induction system
design of the base car. However, even in comparison with some
of the best designed systems, we achieve a significant improvement.
« s
'<
Figure 3 shows a comparison of the attributes of a carburetor
and a Dresserator sonic system. As can be seen, the Dresserator
Inductor has a mass flow metering capability, is an excellent
-------
CARBURETOR-DRESSERATOR FUNCTIONAL COMPARISON
CARBURETOR DRESSERATOR
MASS FLOW CONTROL NONE YES
ATOMIZER POOR EXCELLENT
MIXER POOR EXCELLENT
Figure 3
-------
13
atomizer, and an excellent mixer; whereas, a carburetor provides
only good atomization at high manifold vacuum conditions and is
inherently a poor mixer since it utilizes a butterfly which divides
the flow between a fuel-air stream and an air stream.
Figure 4 shows the comparison of the flow characterisitcs
a carburetor and the Dresserator Inductor plotting manifold vacuum
versus air flow. Each curve represents the point at which the
system can maintain sonic velocity in its throat. Thus, all
manifold- conditions above the lines would indicate where the
system is sonic and below that where it is subsonic. As one
can see with the properly designed Dresserator Inductor, the
system maintains sonic velocity over almost the whole manifold
vacuum range.
Figure 5 shows a Dresserator Inductor test fixture
operating on an engine dynamometer. This fixture was used to
study fuel presentation to the throat for proper atomization.
As can be seen through the plastic side of the unit, the throat
is filled with a cloud of finely atomized fuel.
Application of the Concept
The Dresserator principle can be uti 1 izerU-j.Q .a_y.arJ..ety of
geometric shapes. A Model I has an annular throat containing a
moveable pintle which modulates the throat area. This is shown
in Figure 6. The diffuser is the annular space between the
pintle and the throttle body wall. This geometry is excellent
for use as a valve. It has some limitations when adapted as a
carburetion system, one of these bein:g a problem of idle fuel
» -*
distribution around the very large periphery of the throat.
This peripheral distance can be in the range of seven inches
on our larger units. In addition, the top-opener as shown here,
has limitations on the flow range that it can handle and still
-------
/
/High
/>Flow,
! 1.2!
Flow
0.3"
/
7
AHigh Flow, carb,_
2Bbl. Chev. carb.
50 entrance^,
0.3 " diffuser at 5°
80 to 17° entrance
,/. 1.2" diffuser at 5°
48CFM*= 50-mphat17:1
\
VacuumJrvi:Jl9L
.10
15
-------
-------
-------
17
maintain sonic velocity. It can be designed efficient at the
full open range, or the low-flow range, but does not have the
turndown capability with efficient energy recovery of other
models. A variety of the Model I involves an inverted system
that we call a bottom-opener in which the throat is located
at the smallest end of the pintle. The flow diverges out through
the diffuser as opposed to the converging shape of the flow path
of the top-opener. Characteristics of the bottom-opener are an
improved idle fuel distribution and a vastly improved turndown
ratio. The Model I bottom-opener is excellent as a valve and
is our preferred geometry for a sonic EGR valve.
Figure 7 shows a rectangular shaped unit that we call
a Model II. It consists of two shaped jaws which can be
modulated by sliding apart or by pivoting around a top pivot
point. The latter variant is the model used on the Chevrolet
Nova. Fuel is distributed to this model through a fuel bar
placed in the entrance. The unit has an excellent control over
the area ratio of the diffuser and, thus, has a wide range
efficiency with good idle fuel control.
*
A second rectangular model is shown in Figure 8. It is
called a Model III and utilizes two fixed jaws with a slide in
between to modulate the throat. This unit has excellent idle
fuel control capability as well as an excellent turndown ratio.
It has another quite unique feature: if the entrance slider
wall and its opposite face are kept parallel, the unit has a
constant area ratio in the entrance and, therefore, as long as
the throat is sonic, it has constant depression at any point in
the entrance zone providing an inherent constant depression
metering capability.
These various Inductor models can be mated to a variety
of metering systems. These include:
-------
-------
1. ...-•] 1
awre 5
-------
20
e float fed
« float fed (constant depression)
• electronic float fed
» electronic pressure fed
^ P control
speed density
sonic mass flow
Any of the geometries can be mated with either float-
fed or pressure-fed metering systems. However, it has been our
experience that some of these have attributes more adaptable
to one system than another. If one were to build a float-fed
Dresserator Inductor, we would recommend the Model III, using
its inherent constant depression for the metering. This is the
choice of British Leyland. For a pressure-fed system, we prefer
the Model II since it is a simpler throttle body.
The Model II System
s
The system used on the 1977 Chevrolet Nova is a pivoting
jaw, Model II Inductor with speed density controlled fuel meter-
ing through the use of two fuel injectors feeding fuel above the
throat through a fuel bar. The Inductor system is shown in
Figure 9. This unit is a culmination of a consf^era"bTe"~am6unt
of research and design effort and is an improvement over the
Model II pivoting jaw system as developed by the Ford Motor
Company. Ford designed and built their pivoting jaw sonic carbu-
retor with design features contrary to Dresser design criteria.
These features were:
f
o fuel injection below the?';throat
e no fuel pre-distribution
o unstable diffuser design (stall)
.» high exit velocity - mixture impaction
-------
I \ I V..._J \ .
ADVANCED TECHNOLOGY CENTER
\o
-------
22
a a noticeable foot effort
» poor atomization control at wide-open throttle
By introducing fuel below the throat and without pre-
distribution, the atomizing capability of the Inductor is not
utilized. The diffuser design was such that it had flow
instabilities and a high exit velocity which lead to mixture
impaction on the manifold floor. We have found this to result
in high NOx and reduced fuel economy. The particular diffuser
design also resulted in a noticeable foot effort when coming
off a deceleration and going back onto the throttle. In
addition, the throat width required a throat opening for wide-
open throttle which had a very poor atomization control in
this mode.
Despite these problems, in contest with other systems
at the 1 gpm NOx level with three-way catalyst equipped cars,
Ford obtained better driveability with equivalent emissions
and economy results than with the others examined, namely with
a feedback carburetor and with electronic fuel injection.
Dresser modifications to the Ford system showed a potential
economy gain of greater than 10%.
The Dresser designed system as shown in Figure 9,
incorporates the following design refinements:""
o pulse-dampened fuel metering
o fuel introduction above throat
o increased width and decreased throat opening
o designed out of stall (s.table flow)
o no foot resistance and simplified sealing
» low exit velocity
-------
23
As can be seen in Figure 9, fuel is metered 'through two
fuel injectors feeding opposite sides of a fuel bar shown in the
cut-away. The fuel passages of this fuel bar produce a dampening
of the pulses inherent in a fuel injector. The fuel proceeds
through the downcomers and is pre-distributed across the throat
by a threaded bar place below the downcomer. Fuel is then
atomized as it passes through the throat and the diffuser. The
diffuser incorporates a new design concept: the diffuser is
porous, incorporating a series of holes in each jaw. These have
the function to prevent the supersonic zone from extending further
down the throat than the top row of holes. As the supersonic
zone attempts to come down, it creates a high vacuum which causes
flow from the back side of the jaws into the diffuser, leading
in time to the supersonic flow shocking back to subsonic flow for
the remainder of the diffuser. Since the shock and supersonic
zone is kept high in the throat, there is no tendency for flow
separation. Flow stability is obtained in the basic design as
can be seen in Figure 10 which is one of our design charts that
we have for each of our different models. Shown on this design
chart are iso-unchoke points or vacuums at which level the system
becomes subsonic. The ratio of the throat opening to diffuser
length is plotted on the abscissa and the included angle of the
diffuser is plotted on the ordinate. This angle increases as
the diffuser opens since the system is pivoted from the top. The
graph also contains two lines, one labeled "n^-sJtaJLL!L_and the
other "developed stall". These are obtained from an adaption of
Dr. Kline's work at Stanford University which shows that a
diffuser designed below these lines is very stable but a diffuser
designed above the lines is unstable. Design betv/een the
lines can have flow switching and recirculation, if provoked by
foreign bodies such as by introduction of fuel below the throat
or by tubes extending into the throat, as is the case with the
-------
24
20C
10°
DESIGN CHART
TOP-PIVOT MODEL I
e.
o
o
1.0
Figure 10
-------
25
Ford design. The dark line labeled "Ford dual-jaw" is an operating
line showing where that design was located. The line labeled
"Dresser dual-jaw" is the design used on the Chevrolet and is laid
out to maximize the energy recovery as shown by the iso-unchoke
lines and yet maintain a high stability.
The holes introduced in the jaw eliminates the cause of foot
resistance by preventing a high vacuum zone from being created in
the diffuser. By purposely causing leakage between the front and
back of the diffuser, sealing is only required through the throat
zone and on the back side of the jaws and need not extend below
the top holes. The prevention of the supersonic zone from extend-
ing too far into the diffuser also prevents the jetting of a high
velocity flow from the diffuser and thereby gives a very low exit
velocity with the diffuser running full.
The metering system used with the Model II Dresserator
Inductor is a speed density controlled, dual injector, single-point
injection system. Two injectors feed fuel above the throat of the
*
Dresserator Inductor. The .Inductor is mounted on a single-plane
manifold and no changes have been made to the engine or to the
spark and EGR regimes. The fuel control system is shown schemati-
cally in Figure 11. It incorporates several components from the
Cadillac fuel injection system: namely, the fuel ,pump regulator
injector nozzles, as well as manifold pressure, air temperature,
water temperature, and throttle position sensors. We utilize a
different crankshaft position sensor than on the Cadillac since
this is an experimental system and we wished to be able to study
injector timing. Injector timing, however, was found not to
have any significance in this system. .;.
'*
The electronic control unit is shown schematically in
Figure 12 and is of our own design and fabricated in our lab-
oratory. It is a straightforward speed density system used in
-------
MODEL II - III — FUEL CONTROL SYSTEM — FUEL INJECTION
ENGINE SENSOR PACKAGE
a. MANIFOLD PRESSURE
b. .AIR TEMPERATURE
c. WATER TEMPERATURE
d. CRANKSHAFT POSITION
e. THROTTLE POSITION
FUEL TANK
[SENSOR
IN-PUTS
FUEL RETURN
ELECTRONIC
CONTROL
UNIT
DRESSERATOR UNIT
SIGNAL TO
'INJECTORS
NJ
FUEL INJECTOR
NOZZLES
\
FIXED
PRESSURE
REGULATOR
Figure 11
-------
FUEL/AIR RATIO SET POINT
M.A.P,
COOLANT
TEMP,
COLD START
&
TEMPERATURE
CORRECTIONS
SUMMING
JUNCTION
MANIFOLD ABS,
PRESSURE
AIR TEMP.;:-
MAP
RPM
MASS
(AIR)
PMAN VDISP
R TA,R
VEFF(CoR
ENGINE MAP
(VOLUMETRIC
EFF', COR,)
M
AIR
"(FUEL) ~
F/A(MAIR)
F/A
THROTTLE POSITION'
CRANKSHAFT POSITION-
INJECTOR
DRIVE
NJ
—I
Figure 12
-------
28
an open-loop. A commercial system woul'd use a closed-loop function
in a hybrid system which would operate open-loop and only utilize
the oxygen sensor closed-loop as a calibrating means. Thus, we
avoid the A/F ratio cycling necessary to sense the stoichio-
metric point during most of the operation. This is a-signifi-
cant factor in our results as it helps to maintain o-ur cycle-by-
cycle A/F ratio control which we feel is necessary for high TWC
catalyst efficiency.
Results
The inductor and fuel control system have been operated
and tested in two modes. The first was lean operation at a 1.5
gpm NOx level. Tests were run at DATeC and at the General Motors
facility in Van Nuys. The latter confirm our results with the
exception that we tend to measure a slightly higher NOx level and
a slightly lower economy. However, when compared to the base car
baseline values, we show a consistent economy gain, as did the GM
results. These results are shown in Figure 13. Lean operation
results are run without an air pump and are run without EGR
in the range of 18.5/1 or with EGR in the range of 17.5/1. In
the lean operation, we utilized an old-type, vacuum-operated EGR
valve since we did not have ported signals for control of the
back-pressure EGR. Our economy suffered because~?^r'~th"±s~:
Additional results are shown on Figure 13 utilizing another
device that we are developing called an "Economizer". This is a
very simple, low-cost device which improves economy in the range
of 5-10%. As can be seen with these results, the economy improved
an additional 5% over the baseline economy when the Economizer was
incorporated into the system.
-------
LEAN OPERATION RESULTS
LAB
DATEC
DATEC
GM
GM
BASELINE
.(GM) ^
••
DATEC
GM
HC, GPM
0,22
0,28
0,34
0,33
0,22
0,33
0,52 !!
CO, GPM
1,67
1,96
1,01
0,95
2,35
WITH
1,22
i 0,81
NOX, GPM
1,80
1,67
1,48
1,47
1,22
ECONOMIZER
1,72
1,50
MPG CITY •-
12,70
13,00
13,56
13,51
12,89
13,30
14,15
MPG HWY
17,90
18,10
18,35
17,40
18,3
19,54
VO
Figure 13
-------
30
Stoichiometric operation results are shown in Figure 14,
as run at DATeC, GM, and at the California Air Resources Board.
All laboratories agree on the emissions. The only variation
between labs was on the economy, again ours being lowest. These
results show an economy gain in the range of 10% over the base
car. Results are also shown without an Economizer, again showing
an approximate 0.5 mile per gallon difference. Recently, we found
that we could operate with a single three-way catalyst in place of
the dual catalyst and eliminate the air pump and oxidation catalyst,
achieving even lower NOx levels at the same economy gain. These resul
are shown in Figure 14A along with results without the Economizer.
Development Status
The current development status of the Model II Inductor
system puts it in position to be readily adapted in minimum time
to any size of automobile. There are no fundamental unknowns.
The system has been carefully researched and incorporates the
latest and most up-to-date findings of the Dresserator principle.
The manufacturability has been studied intensively by Ford and the
system found to be easily manufacturable at low cost. In summary,
the current system is:
9 a prototype adaptable to commercial use
o manufacturable
• no critical tolerances
o few moving parts
o low maintenance
• low cost
o adaptable to various engine sizes
A comparison of the Dresserator ..Model II system with other
!• ",
potential systems at various emission levels is shown in Figure 15.
At the 1.5 gpm NOx level or for utilization of the sytem in Europe,
we would recommend a lean operation with a float-fed Inductor.
Under these, conditions, one would gain 10-20% in economy and have
-------
LAB
HC,GPM
STOICHIOMETRIC OPERATION RESULTS
TWC + OXIDATION CATALYSTS
CO,GPH NOX, 6PH MPG CITY
MPG HWY
DATEC
DAT EC
GM
GM
GM
GM"'"'
ARB*
ARB*
BASELINE
(GM)
DATEC
0,29
0,28
0,24
0,30
0,29
0,28
0,29
0,36.
0,22
0,22
2,51
2,83
1.60
2,12
2,17
1,93
2,49
3,42
2,35
WITHOUT E
2,26
0,34
0,33
0,42
0,33
0,33
0,48
0,297
0,32
1,22
CONOMIZER
0,37
13,53
13,60
14,24
14,23
14,11
14,70
14,20
14,10
12,89
13,13
18,6
--
19,02
19,01
--
18,70
18,70
.17,40
—
See Appendix A
Figure 14
-------
LAB
STOICHIOMETRIC OPERATION RESULTS
SINGLE TWC ONLY
HC, GPH CO. GPN NOX. GPN NPG CITY NPG HWY
DATEC
DATEC
0,28
0,36
"2,97
3,12
0,09
0,11
13,63
13,58
WITHOUT ECONOMIZER
DATEC
GM
0,20
0,20
(p,26
2,73
3,10
2,97
0,36
0,29
0,28
12,60
12,75
13,11
Figure 14A
-------
SYSTEM COMPARISONS
SYSTEM
OX. CATALYST
E6R + TWC
EGR + TWC
CRITICAL
STANDARD
0,41 HC
1,5 NOx
HC
1,0 NOx
0,41 HC
0,4 NOx
OPERATION
LEAN
STOICHIOMETRIC
STOICHIOMETRIC
DRESSERATOR
INDUCTOR
FUEL CONTROL
FLOAT FED
PRESSURE +
ELECTRONICS
PRESSURE +
ELECTRONICS
OJ
ECONOMY GAIN
10 - 207o
5 - 15%
5 -
COST
.ADVANTAGE
$80
.$80 - $100
$80 - $100
COMPETITION
AIR, EGR,
OX, CATALYST
F,B, CARB, OR
EFM, EGR, TWC,
AIR, OX, CAT,
FI, FB CARB, OR
EFM, EGR, TWC,
AIR, OX, CAT,
Figure 15
-------
34
an $80 advantage over other systems, mainly because we do not
need an air pump or EGR. At the 1 gpm NOx level, we would operate
at stoichiometric A/F ratio with a three-way catalyst and EGR
using a pressure fed metering system. We anticipated and have
demonstrated an economy gain of 5-15% and would have a consider-
able cost advantage, in the range of $80-$100, over a feedback
carburetor or EFM fuel control system, including EGR three-way
catalyst, air and oxidation catalyst. With our most recent results,
the comparison at 0.4 gpm NOx becomes essentially the same as
thatat the 1 gpm level. In addition, it appears that we have the
potential of meeting the 1 gpm NOx standard without using EGR.
The Model II Inductor system, as described above, does not
fully utilize all of the potential of the Dresser sonic carburetor
concept. As mentioned earlier, the sonic carburetor is a mass
flow metering and measuring device. Thus, the Inductor itself
can be used to measure air flow to the engine rather than computing
it from speed and density. A schematic showing an advance control
system utilizing fully the Dresser concept is shown in Figure 16.
%
Air flow is measured by a throttle position sensor. Flow would be
corrected for air temperature and atmospheric pressure variation.
The atmospheric pressure correction would be by absolute manifold
pressure on start. Once in operation, the pressure correction is
taken care of by the oxygen sensor utilized in -feke—-s-ewi—epen loop
system, EGR would be programmed using the sonic principle in a
sonic EGR valve. However, it is totally independent of the air
fuel system since we have a primary measure on the air mass flow
through the throttle body itself. This scheme provides the
ultimate in simplicity for a control system with a minimum of
variations which can detract from th'e ability of the system to
control and is unique to the Dresserator concept. An additional
benefit is the elimination of the need for extensive engine mapping
lo determine volumetric efficiency.
-------
COOLANT TEMP
M.A.P,
1ROTTLE
POSITION'
\NIFOLD
ABS, PRESS
AIR FLOW
MAP
COLD
START
COR,
AIR TEMP
COR,
AIR TEMP
F/A SET POINT
(.068 - ,067)
M(AIR)
M(FUEL) :
F/A-(M(AiR)>
INJECTOR
DRIVE
CRANKSHAFT Pos,
OXYGEN
SENSOR
OXYGEN
FEEDBACK
SYSTEM
Figure 16
------- |