Summary and Analysis of Comments to the
Notice of Proposed Rulemaking:
"Evaporative Emission Regulation and Test Procedure
for Gasoline-Fueled Heavy-Duty Vehicles"
p p v-l 30; i S 8 0
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Sources
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency

-------
Table of Contents
Paga
I.	introduction	.		 .	i
II.	List of Commenters	ii
III.	issues
A.	Certification Procedure 	 ... 		1
B.	Incomplete Vehicles 		45
C.	Technical Feasibility 		60
D.	Leadtime				70
E.	Costs		83
F.	Test Procedure	92

-------
-1 -
I. Introduction
The Environmental Protection Agency (EPA) published a
Notice of Proposed Rulemaking (NPRM) on Wednesday, April 30,
1S80. The NPRM proposed new regulations for the control of
evaporative emissions from gasoline-fueled heavy-duty vehicles
(HDGs). These proposed regulations included a 3.0 gram per
test (g/test) standard to be implemented for 1S83 and later
model year HDGs. A public hearing on the NPRM was held
Wednesday, June 25, 1980 at the Agency's Motor Vehicle
Emissions Laboratory in Ann Arbor, Michigan.
This document summarizes and analyses the comments
received at the public hearing as well as the final, written
comments ' received after the public hearing. The NPRM,
transcript of the public hearing, manufacturers' final, written
comments, and all other relevant documents have been placed in
the public docket for this rulemaking (Docket #OMSAPC-79-l) .
The public docket can be examined at EPA's Central Docket
Section, W. Tower Lobby Gallery I, 401 M Street, S.W.,
Washington, D.C., 20460.

-------

-i i-

List
of Commenters

1.
Chrysler Corporation
Chrysler
2.
Clayton Manufacturing Company
Clayton
3.
Ford Motor Company
Ford
4.
General Motors Corporation
GM
5.
Griffiths, Russell J.

.6 •
Internatinal Harvester Corporation
IHC
7.
Motor Vehicle Manufacturers Association
MVMA
8.
National Automobile Dealers Association
NADA
9.
National Truck Equipment Association
NTEA
10.
Truck Boay and Equipment Association
TBEA

-------
Issue:
Certification Procedure
Summary of Issue
The certification procedure proposed in the NPRM would
have divided each manufacturer's product .line into evaporative
emission families and, then, subdivide these families into
evaporative emission control systems. The criteria used to
determine evaporative families were selected so that vehicles
in the same family would be expected to emit about the same
.level of evaporative emissions if uncontrolled. To be classed
in the same family, vehicles would have had to be identical
with respect to nominal fuel tank capacity (within a 20 gallon
or 25 percent range, whichever is greater) method of fuel/air
metering (i.e., carburetion vs. fuel injection), and carburetor
fuel bowl volume a lOcc range).
Evaporative emission familes would have been subdivided
into evaporative emission control systems. The criteria used
for this class- ification are parameters which affect the level
of control of evaporative emission. To be classed in the same
control system, vehicles Would have to be identical with
respect to the method of vapor storage, the vapor storage
material, the vapor storage working capacity (within a 20-gram
range), the method of purging stored vapors, and the method of
carbuetor fuel bowl venting during both engine operation and
engine off.
After a manufacturer's product line had been divided into
evaporative families and systems, the Administrator would have
selected one vehicle from each evaporative family-system
combination for testing for evaporative emission data. The
Administrator would also be able to select a maximum of two (2)
additional vehicles within each evaporative emission family for
testing.
Once the emission-data vehicle had been selected, the
manufacturer would build and test the vehicles according to a
full-SHED test procedure similar to the light-duty evaporative
emissions test procedure. (See the issue titled "Test
Procedure" in this document for a detailed discussion) . The
vehicle would be preconditioned over the proposed heavy-duty
chassis driving cycle and then soaked indoors between 10 and 36
hours. After soaking, the fuel tank(s) would be drained and
refilled to 40 percent capacity with chilled test fuel (60°F).
Next, the vehicle would be pushed into the air-tight SHED
(sealed housing for evaporative determination) and' the l-houc
diurnal heat build would begin. During this part of the test,
the fuel would be heated from 60°F to 84°F over a 1 -hour
period. At the end of the hour, the total grams cii
hydrocarbons (HC) emitted would be calculated- from the
concentrations of HC existing before and after the l-hour
diurnal heat builc.

-------
-2-
The next portion of the test consists of the vehicle being
either pushed or driven onto a chassis dynamometer within 1
hour of the- end of the diurnal portion of the test. The
vehicle would then be operated over the heavy-duty chassis
driving cycle. The purpose of this part of the test is
threefold: 1) to warm-up the vehicle for the hot-soak test, 2)
to purge the canisters and 3) to check running losses if
necessary. After being driven over the cycle, the vehicle
would be put back in the SHED (within ten minutes) where it is
kept for one hour. At the end of the hour, the grams of HC
emitted is calculated. This concludes the hotsoak portion of
the test.
The grams of HC emitted during the diurnal heat build are
added to the grams of HC from the hot-soak to arrive at the
total test results. This total plus the d.f. would have to be
less than the emission standard of 3.0 g/test.
Summary of Comments
Comments were received from the four primary HDG
manufacturers (i.e., GM, Ford, Chrysler and IH), MVMA and two
secondary manufacturer trade associations (i.e., Truck Body and
Equipment Association (TBEA) and National Automobile Dealers
Association (NADA)). MVMA, GM, Ford, IH and NADA all commented
that EPA should form a joint Industry/EPA workgroup to develop
a HDG evaporative emission test procedure. These commenters
all suggest that a component test procedure could be developed
which would correlate well with the full-SHED test procedure.
They also suggest that in the interim EPA adopt an "engineering
evaluation" type of certification procedure similar to that
currently in use in California. Chrysler and TBEA commented
that an "engineering evaluation" should be the final form of
the regulation.
GM included a relatively detailed discussion of a
component test procedure that it developed. GM's component
test procedure is divided into seven parts which follow.
1.	Diurnal losses from the fuel tank. Each different
fuel tank would be mounted on a steel frame, placed in a SHED
and filled to 40 percent capacity with chilled test fuel. Then
the fuel would be heated in the same manner as a full-SHED
diurnal test from 60°-84°F. The hydrocarbon (HC) emissions are
calculated and the result would be the uncontrolled level of HC
emissions from that fuel tank.
2.	Storage system working capacity. The first part of
this procedure is designed to equilibrate the storage system.
In GM's case the storage system is a charcoal cannister. The
canister is loaded with gasoline vapors and purged with air a
total of 12 times. The loading consists of vapor being fed
into the canister at a rate of 4 scfh for 10 minutes. Purging

-------
-3-
consists of an air flow of 50 scfh for 20 minutes. The second
part of this procedure establishes the working capacity of the
storage system. The canister is loaded until a breakthrough of
2 grams HC "occurs. Then the canister is weighed. Next, the
canister is purged by a .15" Hg vacuum for 30 minutes. Finally,
the purged canister is weighed. The weight difference between
the loaded canister and the purged canister would be the
storage system working capacity.
3.	Regeneration of storage system. In this procedure
an engine dynamometer test cycle (most probably the transient
test cycle for exhaust emissions is used to purge the
canister. After the canister is saturated to the 2 gram HC
breakthrough level, it is purged with one engine test cycle.
If the weight loss of the canister is greater than the grams of
HC emitted during one carburetor hot-soak test (see below),
then the purging system is deemed to work adequately for
hot-soak emissions. After 3 engine dynamometer cycles, the
weight loss must exceed the sum of the one fuel tank diurnal
and one, one-hour carburetor hot-soak.
4.	Carburetor hot-soak emissions. In this procedure
the carburetor, its mounting gasket, air cleaner and storage
canister are all connected to a steel frame in their relative
positions and placed in a SHED. The carburetor is attached to
a solid aluminum block which rests on an electric heat source.
All elevations, angles and pressures that would occur if the
system were on a vehicle would be maintained. After the
canisters and/or carbon air cleaners had been purged with 3
engine dynamometer test cycles, the electric heater raises the
temperature of the aluminum block which, in turn, raises the
temperature of the fuel in the carburetor fuel bowl. The
temperature profile of the fuel in the bowl would be the same
as that for the fuel bowl on a vehicle during the 1-hour
hot-soak part of the full-SHED test. The HC emissions are
calculated after one hour and the heat source is turned off.
The system is allowed to cool for 12 hours with the SHED open.
During the twelfth hour the HC emissions from the cold
carburetor assembly would be measured. These carburetor
cold-soak emissions represent a measurable part of the
full-SHED diurnal test.
5.	Running losses. During engine operation all
evaporative emissions should be consumed in the engine.
Schematic layouts ana designs of all liquid and vapor lines as
well as gas cap seals and pressure relief valves should
indicate adequate design against leakage. Fuel tank seams and
fuel level gauge seals could be checked by an air pressure loss
test.
6.	Hose permeation (background). A 2 foot length of
fuel hose is filled 75 percent full with test fuel and the ends
are sealed. It is laid horizontally on a surface and weighed

-------
-4-
daily for 2 weeks. The rate of weight loss between the third
and sixth day is converted to an hourly rate pec linear foot of
internal wetted area. This is the cold-soak permeation rate.
GM claims that experience has shown that a good rule of thumb
is 2.5 X cold-soak permeation rate will be the hot-soak
permeation rate. Thus, the "effective length" of liquid fuel
lines is multiplied by 3.5 and the cold-soak permeation rate to
obtain the test result. The "effective length" of a hose is
the length between unsupported ends of the metal tubes or
connectors.
7. Summation of component test data. The final part of
GiM's component test procedure is the manipulation of the above
6 parts to arrive at a total test result. The first step is to
determine the full-SHED diurnal test estimate from the above
component tests. The results of component test £1 {Diurnal
Losses from the Fuel Tank) are added to the results of
component test #4 (Carburetor Hot-Soak Emissions). This total
should be less than the results of component test #2 (Storage
System Working Capacity) . If the summation of #1 and £4 is
greater than #2 then the difference will be used in the
full-SHED diurnal test estimate. Normally, however, the
summation of #1 and M will be considerably less than #2. In
this instance, #1 is to be multiplied by 0.7 percent and the
result will be used in estimating the full-SHED diurnal test.
GM claims that empirical analysis of its data indicates that
activated carbon is only 99.3 percent effective in adsorbing
KC; hence, 0.7 percent will escape. To this result is added
the carburetor cold-soak losses and the cold hose permeation.
The total represents GM's estimate of the diurnal portion of
the full-SHED test.
An estimation of the hot-soak portion of the full-SHED
test is obtained by adding the carburetor hot-soak emissions
(component test #4) plus the hot-soak hose permeation (i.e.,
2.5 X cold hose permeation). The summation of the full-SHED
diurnal test estimate and the full-SHED hot-soak test estimate
will give the final test result which, after compensating for a
deterioration factor (d.f.) would be compared to the standard.
GM tested 2 vehicles (a heavy-duty pick-up and a C-7
series tractor) in a total of 5 configurations. GM tested each
vehicle in an uncontrolled configuration and with a California
control system installed. Additionally, the pick-up was tested
with an activated charcoal ring in the air cleaner. Each
configuration was tested by the full-SHED test and by GM's
component test. GM claims that the results of its testing
(shown in Table 1) indicate good correlation between its
component test procedure and the full-SHED test procedure.
The Agency also received comments on the proposed r.ethcd
of vehicle classification and selection. GM stated that it
fully supports the KDG selection criteria developed by i-'.VKA end

-------
-5-
Table 1
Full-SHED Test vs. GM's Component Test
Diurnal (g/test) Hot-Soak (gpt)
(background	(background Total
included)	included) (g/test).
Pickup Uncontrolled:
Component Test	36.25	9.93	46.18
Full-SHED Test	36.10	10.67	46.77
Pickup w/ 1979 System;
Component Test	.52	1.39	1.91
Full-SHED Test	.52	1.27	1.79
Pickup w/ 1979 System plus
charcoal in Air Cleaner:
Component Test	.53	.96	1.48
Full-SHED Test	.53	.93	1.46
C-7 Tractor Uncontrolled:
Component Test	50.18	15.68	65.86
Full-SHED Test	48.55	25.49	74.04
C-7 Tractor w/ 1980 System:
Component Test
Full-SHED Test
1.23
1.58
2.75
4.80
3 . 98
6 .38

-------
-6-
submitted to EPA on July 10, 197B. That emission-data vehicle
selection scheme is as follows:
1.	Identify each basic type of carburetor design {e.g.,
IV, 2V, 4V).
2.	Combine each carburetor with the highest sales
volume engine.
3.	Match the above combinations with the highest sales
volume configuration for each basic chassis type using the
following selection criteria.
a.	Select only two-axle vehicles.
b.	If the chassis type is offered both as a complete
(with body) and incomplete (chassis or chassis with cab)
vehicle/ select a completed version. If a complete vehicle is
not offered, the chassis shall be tested in the incomplete
configuration.
c.	Select only gasoline-powered chassis types having an
annual sales volume greater than 1000 units.
4.	Select the highest sales volume fuel tank
configuration for each of the above carburetor/engine/chassis
combinations.
5.	If sales of an optional, larger capacity fuel system
exceeds 10 percent in the specific chassis selected in step 3,
a duplicate vehicle incorporating that option rr.ay be tested.
If the MVMA selection procedure is not adopted then GM
stated that only one additional emission-data vehicle should be
selected per evaporative family instead of two as proposed.
This selection would either be an expected highest emitter or a
highest sales volume model. If the first selection for a
family was on a sale volume basis, then the additional
selection would be an expected highest emitter and vice-versa.
GM also suggested that the range of the family determinant of
carburetor fuel bowl volume be expanded from the proposed 10
cc's to 25 cc's. GM claimed that the proposed range is too
small to allow the combination of any two carburetors into the
same family.
GM and Ford commented that the vehicle classification
system should be changed to allow "worst-case" vehicle
testing. For example, one criterion of the proposal's
evaporative emission family definition is that all vehicles
within the same family must have the same fuel tank capacity
(within 20 gallons, or within 25 percent, whichever is
greater). GM stated that this could lead to unnecessary
certification expense. GM claimed that if two vehicles were

-------
-7-
alike in all respects except fuel tank capacity, they would
still be placed in separate families under the proposed
classification scheme. GM claimed that there would be no need
to test both the vehicles since testing of the "worst-case"
{greater fuel tank capacity) vehicle would also determine
compliance for the other vehicle. However, since the two
vehicles would be in separate families each would need to be
tested due to EPA's proposed classification system.
Another general area of comment concerned the
"dieselization" of the heavy-duty vehicle market. GM, Ford and
IH commented that the advantages of the diesel engine over the
gasoline-fueled engine have caused and will continue to cause
the decline of gasoline-fueled engines in heavy-duty vehicles.
GM commented that as dieselization of the heavy-duty vehicle
market increases, the cost burden for evaporative certification
will become disproportionately higher and could force 100
percent dieselization. GM included two magazine articles which
describe increased dieselization. However, GM did not make any
specific predictions as to when and how much dieselization of
the heavy-duty vehicle classes it expects.
Ford's comments on dieselization were more substantive'
than GM's. Ford's discussion on dieselization included Classes
IV-VIII (14,001 lbs. and above GVW) only. The manufacturer
stated that in 1973 gasoline-fueled engines accounted for 62
percent of new, Class IV-VIII sales. By 197S that percentage
had fallen to 41 percent. Ford projected that by 1984 the
percent of gasoline-fueled engines in new, Class IV-VIII
heavy-duty vehicles will be between 16 and 25 percent. Ford
stated that since the number of new, gasoline-fueled,
heavy-duty vehicles over 14,000 lbs. GVW will decline
substantially, the proposed regulation for these classes of
HDGs is an unnecessary cost burden. Ford claimed, therefore,
that Class IV ana above HDGs should be exempt from any
evaporative emission regulation.
IH suggested that EPA split HDGs at the 16,000 lb. GVW
point. The 212,382 units per year below 16,000 lbs. GVW could
be certified by the LDT test procedure while the 156,4.19
non-California units above 16,000 lbs. GVW could be certified
by a component test procedure. IH claims that HDGs above
16,000 GVW will decline substantially due to dieselization
which is accelerated by government regulation of the HDG
industry.
GM claimed that the proposed certification procedure would
cause the manufacturers to test, by an order of magnitude, nore
HDGs relative to the market than they currently test for LDVs
or LDTs. GM stated that this disproportionate burden shoulc be
cor rected.

-------
-8-
GM, Chrysler and MVMA commented that EPA should include an
exception for low sales volume models in the final
regulations.? GM proposed that all HDG evaporative
family-system combinations which are less than 3 percent of
total HDG production be certified by "design analysis,"
provided that all such family-systems combined total less than
10 percent of the manufacturer's total sales of HDGs. Chrysler
suggested that all models with sales of 300 or less be exempt.
MVMA agreed with GM that evaporative emission family-system
combinations whose projected sales volume was below a certain
percent {MVMA did not propose a percent level) should be
certified by "engineering evaluations."
Ford and IH stated that they each introduce new,
heavy-duty vehicle models in September of each year but don't
introduce new, heavy-duty engine models until December or
January. They claimed that it would be an excessive burden to
have to develop and certify their vehicles for both
introduction periods. Ford recommended that this evaporative
emission regulation be applicable to engine Job #1 dates so as
to coincide with exhaust emission requirements. MVMA stated
that this model introduction offset is needed to use up engine
stock effectively.
GM and MVMA requested special certification procedures for
certain special situations. They commented that Special
Equipment Option (SEO) orders are very small volumes and should
be handled by special procedures. SEOs are custom-built
vehicles where the engine/fuel tank combinations are not a
normal part of the production line. Additionally, GM and MVMA
requested exemptions for those vehicles where the ultimate fuel
system is destined to be liquid propane gas even though such
vehicles leave the manufacturer's plant as gasoline-fueled
vehicles. The commenters claim it is unnecessary to have to
charge the customer for an evaporative control system that will
never be used.
IH requested that if these HDG evaporative emission
regulations cause HDG engine compliance testing to be required,
then such engine compliance testing should be done by either
Subpart D or Subpart H at the manufacturer's option. IH
claimed that it would have carried over its engines'
certifications into 1984 as it has been doing. If IH is
required to recertify its engine families according to Subpart
H they would not pass because they were originally certified
under Subpart D. Thus, IH would have to develop its engine
line to comply for the 1S83 model year and again for 1984 when
the transient test for heavy-duty engines becomes effective; an
excessive and untenable burden.

-------
Discussion and Analysis of Comments
Introduction
In general, the commenters did not believe that the
proposed certification procedure was best. They stated that
the cost of that proposed procedure was excessive in relation
to the benefits to be gained. They recommended that EPA adopt
either a component test procedure developed by an Industry/EPA
work group or an "engineering evaluation" type of certification
procedure (or in some instances a mixture of the two). They
claimed these alternative certification programs would be much
less expensive yet would still assure EPA of good control of
KDG evaporative emissions. The commenters also had specific
suggestions to improve the proposed certification procedure.
EPA has incorporated many of these suggestions into its
recommended certification program.
This discussion will first describe and analyze five
different certification scenarios. They are: 1) the proposal
scenario (ie., the certification procedure described in the
Notice of Proposed Rulemaking (NPRM) that was published April
30, 1980), 2) a California-type engineering evaluation, 3) a
component/mini-SKED procedure, 4) a revised proposal scenario
and, 5) an abbreviated certification procedure. Theoretically,
the best certification procedure is the one that gives the
highest benefit-to-cost ratio. However, it is often difficult,
if not impossible, to accurately assess the cost of all the
different aspects of a program. For example, an accurate cost
assessment of research and development efforts under each
scenario would be very laborious. Furthermore, estimating
benefits can be very difficult. The benefits of the Proposal
scenario can be estimated with a fair amount of confidence
since there is considerable experience from light-duty vehicle
evaporative emission control using a similar certification
scenario. But there is very limited data on the emission
levels from a component/mini-SHED certification procedure.
Emission levels might be accurately estimated if large testing
programs were undertaken but such testing programs would
probably cost more than they are worth. Some aspects of this
issue simply do not readily lend themselves to economic
interpretation.	Examples include the question of
manufacturers' equity ana the need for a testing program that
yields a discreet, emissions number for use in long-term,
nationwide, air quality planning.
The following discussion and analysis will use
quantitative benefit and cost analysis when possible. However,
qualitative analysis must also be used to arrive at the
recommendation. The discussion will be divided into four major
parts. The .first section will discribe each of the five
alternative certification scenarios. The next section will
present the advantages and disadvantages of each alternative.

-------
-10-
Th e third section will compare
while the final section will
recommendations for the
certification procedure.
and contrast the alternatives
present the conclusions and
final evaporative emission
Certification Procedures Considered
The Proposal Certification Procedure. The description of
this first alternative was presented above under the section
titled "Summary of Issue." The reader is referred back to that
section for a review of the certification procedure that was
proposed.
A California-Type Engineering Evaluation. Currently, the
state of California's Air Resources Board (CARB) requires that
all HDGs be equipped with evaporative emission control systems
(EECS). Under the CARB regulation manufactures submit schematic
drawings of the different EECSs that they plan to put on their
HDGs. The manufacturers also describe their different HDGs in
regards to parameters expected to affect evaporative
emissions. For example, they submit information on the
different fuel tank capacities and combinations they expect to
sell with their HDGs. The manufacturers then discuss why the
EECSs they have designed should control HDG evaporative
emissions to a level equivalent to that of light-duty. (2.0
g/test is the current light-duty standard.)
One way they can show equivalency is if the EECS is
currently being used on LDTs. If the control system is
essentially the same as one already on LDTs, then CARB will
accept it as long as the manufacturer accounts for any major
differences between the LDTs and the HDGs. For example, the
HDG may be virtually the same as the LDT in regards to
parameters that affect evaporative emissions except for the
total fuel tank capacity. In this case, the manufacturer might
either show that the charcoal canisters used in the LDT EECS
have sufficient working capacity to handle the extra emissions
from the larger fuel tank(s) or he might present a logical
arguement showing that he has increased the canister(s) working
capacity enough to handle the extra emissions.
Under CARB's HDG evaporative program, manufacturers are
not required to do any testing of the EECSs. However, some
manufacturers do limited testing in order to show that their
systems are controlling to a level equivalent to light-duty.
For instance, Ford used a component test in an attempt to show
that their EECSs were providing adequate control of HDG
evaporative emissions. Ford placed the fuel tank(s) and its
evaporative control system (ie., canister(s), tubing and
valves) in a light-duty SHED. Heating blankets were placed
under the fuel tank(s) to heat the fuel from 60°F to 84°F over
a 1-hour period and emissions were calculated. This
represented the diurnal part of the full-SHED test. Next, the

-------
-11-
carburetor was placed in the light-duty SHED along with its
evaporative emisson control system. The fuel in the carburetor
fuel bowl was heated to simulate the temperature profile that
would occur during the hot-soak portion of the full-SHED test
and the emissions were calculated after one hour. Finally, the
results from the two component tests (plus a background factor)
were summed to obtain the total test results. Ford claimed
that this component test correlated well with the full-SHED
test and that since the results of this component test were
quite low (from .7783 grams to 1.098 grams), Ford concluded
that the control systems were doing an excellent job.
After the manufacturers submit their schematics,
discussions, arguements and/or test data, CARB reviews it and
certifies the control systems. This "engineering evaluation"
is intended to encourage the manufacturer to design HDG EECSs
to control evaporative emissions to a level of about 2.0 grams
per SHED test. Actual confirmatory. SHED testing by CARB is not
allowed. The "engineering evaluation" procedure considered in
this discussion of alternate certification procedures will be
this CARB procedure.
Component/Mini-SHED Procedure. All of the commenters
(except Chrysler) stated that the final certification procedure
should have at its core a component/mini-SHED test procedure.
The basic concept of a component/mini-SHED test procedure is
that rather than test an entire vehicle with its associated
evaporative emission control system in a full-sized HDG SHED,
the different subsystems of the control system would be tested
individually with their associated vehicle parts. For example,
the subsystem used to control emissions from the fuel tank(s)
might be tested by itself. The fuel tank(s) and the control
sub-system are attached to a frame and placed in a SHED where
the diurnal test is run. These component tests would be done
in a light-duty SHED or even in a smaller, less expensive SHED
(mini-SHED). Any vehicle's emission level could be determined
by the summation of the test results for that vehicle's
components.
As described in the "Summary of Comments" of this issue,
GM has developed a component/mini-SHED test procedure. To
summarize that procedure, GM suggests dividing the full-SHED
test procedure into seven component tests which are as
follows: 1) diurnal losses from the fuel tank(s), 2) storage
system working capacity, 3) regeneration of the storage system,
4) carburetor hot-soak emissions, 5) carburetor cold-soak
emissions, 6) tunning losses, and 7) hose permeation.
Instead of testing an emission-data vehicle as a whole,
only the above component tests would be run. After determining
that the storage system working capacity and regeneration are
adequate, the diurnal losses, carburetor hot-soak losses,
carburetor cold-soak losses, and hose permeation are summed.

-------
-12-
Th is result would then be compared to the standard (which is
based on the full-SHED test) to determine compliance. (Running
losses should only rarely occur and are checked by design
analysis and pressure checks.)
Ford has also developed a component/mini-SHED procedure.
Ford developed it in early 1978 for the purpose of complying
with CARB's HDG evaporative emission regulation. It is simpler
than GM's and consists of only two parts: the diurnal test for
fuel tank(s) loses and the hot-soak test for carburetor losses.
The vehicle classification (into families and systems) and
emission-data vehicle selection procedures would remain the
same as they were under the proposal certification procedure.
GM commented that only one additional emission-data vehicle per
evaporative emission family should be tested (instead of two)
and GM and Ford both suggested that the definition of a family
should be changed to allow the testing of "worst case" vehicles
only. The EPA staff believes that since there is a lack of
data correlating a component/mini-SHED test procedure to the
full-SHED test procedure, the increased certification testing
resulting from retention of the proposed family-system
definitions will be somewhat compensatory. Therefore, the
vehicle classification and selection procedures for this
alternative certification procedure will be the same as the
Proposal. However, the advantages and disadvantages of an
alternative vehicle classification and selection scheme will be
discussed in the next section.
The Revised Proposal Certification Procedure. In an
effort to reduce the certification burden, the staff has
reconsidered the proposed certification program. The staff
believes that the following revised certification program would
substantially reduce the time and cost of HDG evaporative
emission certification while retaining most, if not all, of the
air quality benefits of the originally proposed program. This
revision has evolved from written comments, oral comments at
the public hearings, conversations, with the California Air
Resources Board (CARB) and conversations with manufacturers.
It is a combination of the proposed certification plan, the
California-type	engineering	evaluation	and	the
component/mini-SHED testing concept.
The staff believes that Class VII (26,001-33,000 lbs.
Gross Vehicle Weight (GVW)) HDGs and Class VIII (33,001 lbs.
and above GVW) HDGs should be exempt from evaporative emission
testing. These two classes of HDGs would still require
evaporative emission control systems but certification would be
based on an engineering evaluation procedure similar to CAR3's.
As discussed in Chapter III of the "Regulatory Support
Document" for this rulemaking (see the docket), these two
classes of HDGs represent a small portion of new HDGs sold each
year. In 1979 , there were only about 19,000 Class VII HDGs

-------
-13-
sold. In 1989/ EPA projects that new Class VII vehicles will
be virtually all diesel-powered. Since this regulation will be
implemented in 1985, there will be four model years (1985,
1986, 1987, and .1988) that Class VII HDGs will still be sold.
EPA expects that the total number of new, Class VII, heavy-duty
gasoline-fueled vehicles sold during model years 1985-1988
inclusive will be approximately 10,800. This represents less
than 1 percent of all HDGs sold during this period.
EPA projects that there will be virtually no new sales of
Class VIII HDGs for model years 1984 and beyond. Thus,
exempting this class of HDGs from certification testing
requirements should have no impact on air quality.
Since sales of new Class VIII HDGs are expected to be
nonexistent in 1985 ana sales of new Class VII HDGs should
virtually disappear by 1989, the certification of the
evaporative emission control systems for the few new Class VII
and VIII HDGs that are produced should be as inexpensive as
possible. The engineering evaluation procedure would assure
the Agency that these vehicles have control systems that are
comparable to the ones found on Class VI (19,501-26,000 lbs.
GVW) HDGs. If the maximum carburetor fuel bowl capacity and/or
maximum fuel tank(s) capacity offered by a manufacturer (HDGs
only) is only offered on a Class VII or VIII HDG, then the
manufacturer would have to insure that the vapor storage
device(s) has sufficient working capacity to store the
evaporative emissions. The manufacturers would submit
schematics of components and designs showing that the control
systems for their Class VII and VIII HDGs are as effective as
those on their Class VI and lighter HDGs.
The revised certification procedure for Class VI and
lighter HDGs is a combination of the proposed full-SHED
procedure and the component/mini-SHED procedure. This revised
certification program would test an entire vehicle according to
the f ull-SHED procedure but the vehicle tested would be a
"worst case" vehicle. "Worst case" components would be
selected, installed on a "worst case" vehicle, and tested.
Less than "worst case" vehicle components would not need to be
tested thus saving, development and certification effort.
Under this revised certification program a manufacturer's
product line would ' be divided into evaporative emission
families and then those familes would be subdivided into
evaporative emission control systems. Family differentiation
would be very simple. All vehicles which have the same method
of fuel/air metering (i.e., fuel injection vs. carburetion) and
have the same carburetor fuel bowl volume (within a 10 cc range
if applicable) would be an evaporative emission family. These
evaporative family criteria are the same as those in the
proposal except that the proposal included total fuel tank
capacity (within a 20 gallon range or 25 percent, whichever is

-------
-14-
greater) as a family determinant. As will be discussed later,
total fuel tank capacity is still considered in this revised
certification program.
To be classed in the same evaporative emission control
system, vehicles would have to be identical with respect to the
method of vapor storage, the method of carburetor sealing, the
method of air cleaner sealing, the vapor storage working
capacity (within 20 grams), the number of storage devices, the
method of purging stored vapors, the specifications of the
purge system, the method of venting the carburetor during both
engine off and engine operation, and the liquid fuel hose
material. This classification scheme for evaporative emission
control systems is essentially the same as the proposed scheme
except more detailed and comprehensive. The objective of this
classification scheme is to include all parameters which may be
expected to significantly affect evaporative emission control.
Once a manufacturer's product line had been divided into
evaporative emission family-system combinations, the
manufacturer would select a "worst case" vehicle for
emission-data testing from each family-system combination. The
criteria for selecting a "worst case" fall into two
categories; 1) criteria that are model-dependent and 2]
criteria that are not model-dependent. A criterion would be
defined as model-dependent if, in selecting the "worst case"
situation for that criterion, the only vehicles considered are
those of a particular model or models rather than all of the
vehicles in the entire family-system combination. This
definition will become clearer in the discussion ana examples
that follow.
The first criterion for the selection of a worst case
emission-date vehicle is total fuel tank capacity. This
criterion is not model dependent, that is, the manufacturer
would consider ail vehicles (and options for those vehicles) of
the entire family-system combination and then select that fuel
tank or fuel tank combination which has the largest total
capacity. The next criterion, length of nonmetal fuel hose, is
also not model-dependent. The manufacturer would consider all
vehicles of the family-system combination and choose the
maximum length of nonmetal fuel hose. Only fuel hose that
carries liquid fuel would be considered since the permeation
rate of fuel vapors through fuel hose is considered negligible.
The final criterion which is not model-dependent is the
maximum exposure of the nonmetal fuel hose to heat from the
exhaust system and other sources. The manufacturer wo^ld
review the fuel hose routing on all cf the models of r_he
family-system combination. Be would then se". ect that routing
which exposes the fuel hose to the greatest amount of
Although the manufacturer would use his engineering judgment to
decide which routing is "worst case," he would still submit to

-------
-15-
EPA the fuel hose routing diagrams for all cf the different
vehicles within the familysystem combination showing heat
sources as well as the fuel hose routing. The Agency would
reserve the option of selecting a different fuel hose routing
if it considers such routing to be exposed to more heat than
the routing selected by the manufacturer.
In summary, the manufacturer would consider all vehicles
of each evaporative emission family-evaporative emission
control system which he wishes to market. He would determine
the maximum fuel tank capacity, the maximum length of nonmetal
fuel hose and the fuel hose routing which exposes the fuel hose
to the most heat. These three criteria would be incorporated
onto the emission-data vehicle once it is built. The following
discussion explains the selection of the emission-data vehicles.
After having selected the above criteria which are not
model-dependent, the manufacturer would next select the "worst
case" model on which to install those selections. First, the
manufacturer would list all models of the evaporative
family-system onto which the previously selected largest
maximum fuel tank or fuel tank combination can be installed.
Models to be considered are not limited to only those models
where the manufacturer would normally offer the previously
selected largest fuel tank or fuel tank combination. Any model
that can reasonably be' expected to accomodate the selected fuel
tank or fuel tank combination shall be listed. Reasonableness
includes consideration of safety and testing facility
limitations. Furthermore, in considering whether or not a fuel
tank(s) can be installed on a model (for the purposes of this
regulation) the fuel tank(s) proximity to exhaust system heat
must be approximately maintained.
From this list of models the manufacturer would select
that model having the highest Gross Vehicle Weight Rating
(GVWR) up to 26,000 lbs. This will be the model of the
emission-data vehicle build. If more than one model fits the
above conditions, then the selection of the model of
emission-data vehicle would be based on first/ the proximity of
the fuel tank(s) to exhaust system heat and secondly, the
engine compar t merit size, (i.e., the model with the smallest
engine compartment would be considered "worst case").
After the "worst case" model had been selected, the
manufacturer would install the previously selected "worst case"
fuel tank(s) on a vehicle of that model. He would also install
the previously selected "worst case" length of fuel hosing in
accordance with the previously selected route that exposed the
fuel hosing to the most heat. The resulting vehicle would be
the emission-data vehicle for that evaporative family-system.
Once the emission-data vehicle had been selected . ~nd
built, it wcula be tested according to the proposed full-S'-'ED

-------
-16-
test procedure (except for changes discussed in this document
under the issue entitled "Test Procedure"). Thus, although the
test procedure and standard would not change substantially,
this revised vehicle classification and selection procedure
would substantially reduce the burden of certification by
establishing fewer evaporative family-systems and, thereby,
would require less developmental and certification testing
effort. This reduction of the certification burden will be
discussed in detail in the next section.
After the manufacturer had tested its product line it
would submit the test data to EPA. EPA would then review the
data. The rest of the certification procedure would parallel
the light-duty procedure in that EPA could immediately issue a
certificate, or could request additional data or could do
confirmatory testing. Each manufacturer would have to
establish a deterioration factor (d.f.) as proposed in the
NPRM. Each manufacturer would be free to develop its own test
procedure for deterioration (which would be submitted with its
Part I application) but that test procedure would have to
include an evaluation of the effects of vibration, the vapor
load-purge cycling of the vapor control system, and the aging
effects of heat and ozone.
Abbreviated Certification. This certification procedure
is the simplest of the five alternatives considered.
Abbreviated Certification would establish a 3.0/4.0 g/test
standard and the test procedure discussed earlier under the
Proposal and Revised Proposal certification alternatives. (For
a detailed description of this test procedure, see the issue
titled "Test Procedure" in this document.) Also, the
evaporative emission family-control system vehicle
classification scheme discussed under the Revised Proposal
alternative would be used for this alternative. The difference
is that manufacturers would not need to submit any test data or
engineering evaluation to show that their evaporative emission
control system met the standard. Instead, manufacturers would
be required to submit a written statement that their HDGs would
meet the appropriate standard if tested in the case of Class
113 through VI HDGs and that their control systems are designed
to meet a 4.0 g/test standard for Class VII and VIII HDGs.
Although EPA would retain the option to deny certification
and/or to do confirmatory testing, we would not expect to
actually do any since we are confident that the manufacturers
can easily meet the 3.0/4,0 g/test standard. Given that; 1)
the standards ace not as stringent as proposed, 2) the
technology is straightforward and 3) there exists a great
amount of directly applicable experience from LDT evaporative
emission control, we believe that the manufacturers will be
able to install their systems on HDGs and be fairly confident
of meeting the standards. Furthermore, while the manufacturers
would still be required to develop their own procedures for
determining d.f.s and such procedures would still need to

-------
-17-
account for the effects of vibration/ ozone, heat and vapor
load-purge cycling, they would not be required to submit the
d.f. test procedures or the established d.f.s but would
themselves apply the d.f.s to their test data. Their
statements of compliance would include the d.f. calculation.
Negative d.f.s would not be allowed.
Advantages and Disadvantages of the Alternatives
In order to compare the different certification
alternatives, the level of emission control will be expressed
in terms of a full-SHED test even though the California-type
Engineering Evaluation and the component/mini-SHED procedures
do not incorporate that test procedure. The estimated level of
control from these two procedures will be expressed in terms of
a full-SHED test.
Also, the cost estimates included in this analysis were
based on an implementation date of the start of the 1984 model
year. Since those estimates were made the implementation date
has been changed to the start of the 1985 model year. However,
since these cost estimates do not change significantly with the
one-year delay in implementation, EPA has not recalculated them.
Since the standard in this Final Rule is a split standard
(3.0 g/test for HDGs with GVWs .14,000 lbs. and below and 4.0
g/test for HDGs with GVWs greater than 14,000 lbs), we have
calculated a single number based on the expected future sales
split between these two groups of HDGs. Our projected sales
split indicates that 53.3 percent of future HDG sales will be
vehicles with GVWs less than 14,000 lbs. which leaves 46.7
percent greater than 14,000 lbs. GVW. By applying these
percentages to the appropriate part of the split standard we
arrive at a single, combined sales-weighted standard of 3.47
g/test. This number will be used through out this discussion
to estimate certification control levels.
The Proposal Certification Procedure. Of the five
certification alternatives considered, the procedure proposed
in the NPRM would control HDG evaporative emissions by the
greatest degree. This procedure would control new HDG
evaporative emissions to a level below the standard of 3.0/4.0
grams per full-SHED test (g/test). Manufacturers must account
for test-to-test, lab-to-lab, and production variability. They
must provide a margin for the deterioration factor (af) and for
their own protection. For model year 1981, light-duty trucks
(LDTs) had to meet a 2.0 g/test evaporative emission standard.
A review of the 1981 LDT certification data shows that the
actual average emission .level was 1.19 g/test. This is 41).5
percent below the standard. We will assume that the average
certification value for HDGs will also be 40.5 percent below
the actual standard. Thus, the application of the above LDT
ratio	(1.19 g/test:2.0 g/test)	to the HDG

-------
-18-
standard yields a control level of 2.06 g/test for new
vehicles. This level of control will be compared to the
control levels of the other certification procedures as part of
the process of selecting the best certification procedure.
Not only is the level of control the greatest for the
proposal certification procedure, but the confidence in this
level of control is highest too. In controlling a relatively
minor mobile source of hydrocarbons such as HDG evaporative
emissions, the in-use level of emissions must be estimated from
the certification level because little in-use testing is done.
In-use emission levels are used in projections of future air
quality. These projections can be on a local or regional scale
and are important for planning future control strategies. By
studying these projections of future air quality, planners
determine which pollution sources, if any, should be controlled
and to what level. For light-duty vehicles (LDVs), evaporative
emission factor programs ana extensive in-use testing are used
to obtain reliable in-use emission levels. But since HDG
evaporative emissions are fairly small, the expense for an
in-use testing program will likely be prohibitive. Thus, in-use
emission levels will have to be estimated from certification
data. The proposal procedure with its associated full-SHED
test procedure will give hard numbers from which in-use
emission levels can be estimated. Furthermore, since this
certification procedure requires the greatest number of
emission-data vehicles to be tested, the confidence in the
emission levels thus obtained is greater than the other
certification alternatives.
Another advantage of this certification procedure is that
the full-SHED test procedure is well known. This test
procedure has been used for evaporative certification of LDVs
for many years. All four HDG manufacturers and EPA are very
familiar with it. In implementing this test procedure for HDG
evaporative emission certification, a minimum of effort would
be needed to initiate the manufacturers and EPA as to how to
run the test, how to interpret the data and what kind of
problems to expect. The component/mini-SHED and the
engineering evaluation certification procedures would require
substantially more learning and debugging effort to put in
place.
Another advantage of the Proposal certification procedure
is its objectivity which is important in a certification
program for two reasons. First, EPA must, for obvious reasons,
treat all manufacturers equally. Therefore, a certification
procedure should have minimal subjectivity since therein lies
the potential for inequity. Two EPA reviewers could analyze
the same application and make different determinations on
subjective portions. (Due to the large number of different
vehicles involved ana due to the fact that future vehicle
configurations and technology are difficult to predict, sc,~e

-------
-13-
measure of subjectivity is usually necessary to maintain
flexibility.) The second reason that objectivity is important
is that EPA has been attempting to move towards manufacturer
self-certification. This wi.ll reduce required government
effort, give manufacturers better control of their
certification programs, and make certification more cost
effective. In a self-certification program, subjectivity must
be kept to a minimum not only to ensure manufacturer equity but
also to assure EPA that control of emissions is not being
compromised by manufacturer interpretation.
The vehicle classification and selection criteria for the
five certification procedures considered are about the same
with repect to objectivity. The criteria for vehicle
classification are quite specific. The criteria for
emission-data vehicle selection do involve analysis and
judgment on EPA's part. For example, under the proposal
procedure, EPA would have the option of selecting up to two
additional emission-data vehicles from each evaporative
emission family after one emission-data vehicle per
family-system had been selected. The proposal intentionally
leaves open the method that EPA would use to select these two
additional emission data vehicles. EPA might choose one or
both on the basis of highest expected sales or on the basis of
highest expected emissions. All of the certification
procedures are potentially subjective in this regard. Vehicle
selection will be addressed further under the Revised Proposal
certification procedure.
Beyond vehicle classification and selection, the test
procedure segment of a certification procedure will have
varying degrees of objectivity. The full-SHED test procedure
is well defined and yields a discrete number which can be
directly compared to a standard. While some test-to-test
variability still exists it is small due to the experience
gained from years of attempting to reduce such variability.
The major disadvantage of the Proposal certification
procedure is cost. Of the five alternative procedures
considered, it is the most costly. The manufacturers included
cost estimates for the proposal procedure in their final,
written comments. These estimates are presented and discussed
as "Issue F: Cost" in the "Summary and Analysis of Comments"
to this rulemaking (in this document). The cost of this HDG
evaporative emission rulemaking can be divided as follows: I)
testing facility and equipment costs, 2) research ind
development costs, 3) certification costs, and 4) control
hardware costs. We have assumed that the control hardware
costs will be about the same for all four certification
alternatives, and therefore, need not be addressed for the
purpose of comparing these alternatives. As shown in the
"Summary and Analysis of Comments," the manufacturer
estimated total cost- for testing facilities and equipment
f

-------
-20-
research and development and certification under the proposal
procedure is about $25M.
Another disadvantage to the Proposal certification
procedure is that it would require the greatest amount of
leadtime of the four alternatives. As discussed under the
issue entitled "Leadtime" (in this document), GM and Ford
project that 34 and 30 months, respectively, would be required
from the date of publication for implementation. This time
would be needed to build new facilities, to install test
equipment, and to do development and certification testing.
The background or nonfuel emissions problem is worst with
the proposal certification alternatives. This problem is
discussed in detail under the issue entitled "Technical
Feasibility" (in this document). To briefly summarize that
discussion, new vehicles emit evaporative emissions from
sources such as new paint, sound deadeners, lubricants, and
tires. The level of evaporative emissions from these sources
decreases rapidly over time and, after about two months,
stabilizes. Manufacturers are allowed to age emission-data
vehicles both naturally and artificially (i.e., by baking or
driving the vehicle) to stabilize background emissions before
testing. Furthermore, manufacturers are allowed to omit these
sources of background emissions if they can. The LDV, LDT and
HDG evaporative emission standards account for the stabilized
level of background evaporative emissions. The biggest problem
is that Selective Enforcement Audits (SEAs), which attempt to
test production vehicles, become very cumbersome because of
these background emissions. It is impractical to select a
vehicle for an SEA and then have to wait a month or two for the
background emissions to stabilize before testing that vehicle.
The manner in which the California-type engineering evaluation
or the component/mini-SHED certification procedures could
alleviate this problem will be discussed below.
The Caiifornia-Type Engineering Evaluation
The greatest advantage of this certification procedure
over the other procedures is its cost. This procedure is the
least expensive of the four alternatives considered. While the
other procedures would require the manufacturers to build new
facilities and/or purchase testing equipment, this procedure
would only require that the manufacturers show by engineering
evaluation that its evaporative emission control systems w ill
control emissions to a level equivalent to LDTs. The other
procedures would require some development effort with its
associated testing expense. This procedure would allow the
manufacturer to simply install its LDT control systems on its
HDGs as long as it accounts for any major differences between
the vehicle classes. The major expense (besides the hardware
costs) of this certification procedure would be the man-hours
involved in drawing up schematics of control systems and

-------
-21-
components and in presenting evaluations that attempt to shew
control systems to be adequate.
Estimates of the cost of an "engineering evaluation"
certification procedure are difficult because of the lack cf
experience with such a procedure and because the expenses
involved are difficult to quantify. For example, there may
well be a substantial amount of paperwork involved but the
number of hours is very difficult to oetermine. Although
conceivably the manufacturers would not have to go any testing
under this certification procedure, it is probable that some
component testing would be done to generate test data in
support of the engineering evaluations. If we assume that the
two smaller HDG manufacturers (Chrysler and IH) purchase one
mini-SHED each (with associated analysing ana recording
instruments) and the two larger manufacturers (GM and Ford)
purchase two mini-SHEDs apiece, then total equipment costs ~fcr
this certification procedure would be 6 x $30,000 = $180,000.
This certification procedure would cause somewhat more
paperwork than the other procedures and there will probably be
some limited amount of development testing that the
manufacturers will do. We will assume these costs total about
$320,000 thereby bringing the total cost of this certification
procedure (excluding hardware) to about $Q.5M.
Another advantage to this procedure is that leadtime is
minimal. There is no question that the manufacturers could
install evaporative emission control systems on their 1984
HDGs. The manufacturers have indicated as much.
This California-type engineering evaluation v»ould reduce
the incomplete vehicle problem greatly. The primary
manufacturer would certify only those vehicles which he
actually produces. The secondary manufacturer would submit its
own engineering evaluation to EPA showing that the additions or
modifications it mace to the original vehicle do not cause that
vehicle to emit evaporative emissions at levels higher than
LDTs. For example, if the secondary manufacturer aads fuel
tanks to an incomplete vehicle, then it would attempt to show,
through an engineering evaluation, that the charcoal
canister(s) have sufficient capacity to handle the evaporative
emissions from those fuel tanks. Since no testing is requited
ana the primary manufacturer would still supply a complete
control system, the burcen on the secondary manufacturer woulc
be " limited to the paperwork neeoed for the engineering
evaluation.

-------
-22-
The background emissions problem disappears under this
certification procedure. Since there is no testing required,
actual emissions (background or otherwise) are never measured
and, therefore, do not pose a problem. However, SEAs and NCPs
are not possible under this certification procedure since
actual test numbers are required for such programs.
A major disadvantage to this certification procedure is
that the level of control would probably be the worst of the
five alternatives considered. California has required
evaporative emission controls on HDGs since 1978. The actual
emission levels (in terms of a full-SHED test) of these
vehicles was previously unknown because the California
procedure does not require the manufacturer to test vehicles
nor does it allow CARS to test vehicles for compliance. In
their final written comments to this proposed rulemaking
several manufacturers submitted data from full-SHED tests of
HDGs with certified California evaporative emission control
systems. From these test results we have estimated what the
level of evaporative emissions from HDGs would be under a
California-type engineering evaluation certification procedure.
Ford undertook the most ambitious testing program. It ran
a total of 126 full-SHED tests which were split among four
trucks. Each truck was tested under various prep cycles,
inertia weights and dynamometer horsepower settings. Only
those valid tests where the prep cycle was the proposed
heavy-duty driving cycle and the control system was the
California system were included in the following average
because it is our intent to evaluate only the level of
emissions from California control systems when tested by the
proposed, full-SHED test procedure. The average emission level
of the 47 tests which meet the above criteria is 3.35 g/test.
GM tested two HDGs with California control systems. Since
GM does not have a heavy-duty chassis dynamometer, it used a
road-coute for the preconditioning and driving cycle parts of
the test. GM patterned the road-route after the heavy-duty
driving schedule in order to simulate as close as possible the
proposed test procedure. The results of the full-SHED testing
of the two trucks averaged 4.09 g/test. GM SHED tested two
HDGs in early 1978 but those results (close to the above
average but somewhat higher) cannot be included here because;
1) the vehicles employed old technology, 2) the driving cycle
used is not known, and 3) the dynamometer limitations prevented
testing of realistic horsepowers.
IH tested three KDGs for evaporative emissions and
submitted the results in its final comments. IH used a ^ H E D
for testing, however, since IH did not have a heavy-cuty
chassis dynamometer it used a road-route of 8.6 miles with
stop, starts, accels, decels and cruises. The use of such a
road-route makes the test data somewhat questionable since Icac

-------
-23-
factors ana purge rates are not the same as if the vehicle were
driven over the heavy-dut^ driving c^cle. Two of the three
KDGs incorporate 1980 California evaporative emission control
systems. The other HDG had a 197S California control system
but this system was modified into a 1980 system by the addition
of a charcoal ring in the air cleaner. IH added new tank
gaskets and screws to two of the vehicles. Tests were
conducted before and after this modification. Since the tests
without the new gaskets and screws best represent the
California control system, only those tests are included here.
IH's average emission level was 4.84 g/test.
The average of the emission levels of Fore's California
control system (3.35 g/test), GM's California control system
(4.09 g/test), ano IH's California control system (4.84 g/test)
equals 4.09 g/test. This is 2.03 g/test (or 99 percent) higher
than the projected emission level (2.06 g/test) unc.tr the
proposal certification procedure.
Another disadvantage of the California-t^pe engineering
evaluation as a certification alternative is the large amount
of subjectivity involved. It would be difficult for EPA to
develop objective guidelines for assessing the manufacturer's
engineering evaluations. As the California experience has
shown, each manufacturer would probably have its own method of
evaluating the adequacy of its own control systems. EPA would
have to assess each manufacturer's submittal differently. EPA
reviewers would have different levels of expertise in different
areas and they might analyse different kinds of data (e.g.,
statistically treated data, raw data, narrative data) in
different ways. The chances that one manufacturer might get a
less critical review than another or that an inferior control
system might slip bi EPA while a better one might be rejected
exist.
A third disadvantage of this certification alternative is
that air quality planners woulo have less accurate numbers to
work with when considering in-use HDG evaporative emission
levels. Since the likelihood of in-use testing programs for
HDG evaporative emissions is small, in-use emission levels >ill
have to be estimates from certification emission levels. if
there is no requirec testing for certification, then
certification emission levels would often be only estimates.
Thus, the confidence in the HDG evaporative numbers usee for
future air quality projections would be reduced.
A final cisacvantage to the California-t}pe engineering
evaluation certification alternative is that the possibility ci
HDG evaporative SLAs is reduced because full-SHE3 testing is
not used under this option.

-------
A Component/Hini-SHED Certification Procedure
This alternative, as well as the next (i.e., the Revised
Proposal procedure), is a compromise between that alternative
which costs the least and supplies the least benefits (the
California-type Engineering Evaluation) and that alternative
which costs the most and supplies the most benefits (the
Proposal procedure). The cost of the component/mini-SHED
certification alternative is moderate. Manufacturers will have
to purchase mini-SHEDs and they will have to allocate or build
facility space for testing. The amount of research and
development effort required under this alternative would, most
likely, be somewhat less than under the Proposal alternative.
However, the testing effort would probably be more since the
manufacturer would have to do perhaps three or four
component/mini-SHED tests to simulate one full-SHED test.
Thus, if we assume the reduction in research and development
effort would be offset by the increase in man-houts required to
do testing (both development and certification), and if we
assume the control system hardware cost is about the same as
the proposal alternative, then the main difference in cost
between the component/mini-SHED alternative and the Proposal
alternative would be the cost of testing facilities and
equipment.
Of the. $25M (not including hardware) estimated for the
Proposal certification procedure approximately $9M was
estimated for certification, research and development.
Therefore, $9M will be estimated for these efforts under the
component/mini-SHED alternative as well. We will assume that
GK and Ford will purchase 3 mini-SHEDs each and Chrysler and IH
will buy 2 each. Thus, if each mini-SHED (and associated
analyzing equipment) costs $30K, then the total cost would be
$300K. GM claims it needs 2 new computers ($150K), a canister
equilibrator ($63K), and various miscellaneous equipment ($20K)
for an additional equipment cost of $233K. If we allow $50K
apiece in additional equipment costs for Ford, Chrysler and IH,
total equipment costs would equal £683K.
The space needed to install equipment and to store
vehicles and vehicle parts must also be included in the cost of
this alternative. If we allow 100 ft2 to park two vehicles,
500 ft2 to install four mini-SHEDs, 500f c2 for storage and 500
ft2 for maneuvering, then total space cost for GM and Ford
would be about $375K each (@$150/ft2). For Chrysler and IH we
will assume 2000 ft2 each for a cost of S300K each. Thus, the
total cost for space is estimated to be $1.35M. When equipment
costs, facility space costs, certification testing costs ana
development ' costs are summed, the resultant total cost
(excluding hardware) for this component/mini-SKED certification
alternative is estimated to be $11M.

-------
-25-
Another advantage of the component/mini-SHED certification
alternative is that the problem of background or nonfuel
emissions from full-SHED testing would be lessened
considerably. Background emission sources include tires,
vehicle paint, lubricants, sound deadeners and sealers. Most
such sources would never enter the mini-SHED. If the whole
engine was placed in the mini-SHED there might be some nonfuel
HC emissions from lubricants or paint but these sources could
easily be minimized or eliminated. Likewise, if the fuel
tank(s) have been painted some nonfuel HC emissions might be
emitted but baking or eliminating the paint would reduce or
eliminate these emissions.
SEAs become a definite possibility under this alternative
only if an industry-wide component/mini-SHED test procedure is
developed. Since it seems highly unlikely that EPA would be
willing to expend the time or resources to develop a good
component/ mini-SHED test procedure that correlates well with
the full-SHED procedure and is repeatable, SEAs under this
alternative are improbable. If SEAs are not a part of the
final rule, then, of course, NCPs would not be either.
Finally, the incomplete vehicle problem would be
significantly moderated under this component/mini-SHED
certification alternative. EPA would establish evaporative
emission family-system parameters for which the primary
manufacturers would set limits. If the secondary manufacturer
stayed within these limits when making modifications or
additions to a vehicle, then the certificate of conformity
would remain valid. If the secondary manufacturers wished to
exceed the limits of one or more of the parameters it would
test that parameter(s) in a mini-SHED or have a primary
manufacturer do such testing. This would be an inexpensive ana
quick thing to do. The burden placed on the primary
manufacturer would be reduced in comparison to the proposal
procedure because the risk involved (concerning customer
satisfaction) if the limits were not extensive enough would be
substantially less.
The biggest disadvantage of the component/mini-SHED
cert ificat ion alternative is that the actual level of
evaporative emissions would not be known as accurately as with
the full-SHED procedure. Both GM and Ford claimed that their
component/miniSHED test procedures correlated very well with
the full-SHED test procedure. However, a deficiency of data,
crucial inconsistencies in the manufacturers' data, and
concerns with the procedures themselves cast doubt on these
claims of full-SHED test equality.
As discussed earlier, Ford's component test procedure
consists of two parts: 1) diurnal vapor loss and 2) hot-soak
vapor loss. Ford claimed that the summation of these two tests
plus the addition of a background emissions correction factor

-------
-26-
gave a total test result which was almost identical to the
results of the full-SHED test of a .1978 LDT with the same fuel
system. The component test result was 1.38 grams and the
full-SKED test result was 1.46 grams. This appears to be
fairly good correlation. However, later test results do not
uphold this correlation.
Ford developed this component test procedure to support
the engineering evaluation of its HDG evaporative emission
control systems for California certification. In its
application for certification of its 1980 model year California
HDG evaporative control systems, Ford included its component
test results for a system designed to control a vehicle with
dual 75-gall on fuel tanks and a 2 V carburetor with a bowl
volume of 134 cc. The average of the three tests submitted, was
1.098 grams. If we include Ford's estimation of stabilized,
HDG background emissions of .50 grams then the total test
result equals 1.60 grams.
For their final written comments to this rulemaking, Ford
tested a vehicle with a 2V carburetor and dual 75-gallon fuel
tanks according to the full-SHED test procedure. While we are
not sure that the carburetor and fuel tanks are exactly the
same as the ones tested earlier by the component procedure, the
differences between the fuel system tested for 1980 model year
California certification and the fuel system tested for this
rulemaking are probably minimal. Additionally, this vehicle
was equipped with Ford's California evaporative emission
control system. However, the evaporative emissions from the
full-SHED test procedure averaged 116 percent higher (3.46
g/test: 1.60 g/test) than the component test results. This is
not good correlation between the two test procedures.
Also for 1980 California HDG certification Ford tested a
fuel system consisting of a 4V carburetor with a fuel bowl
volume of 224 cc and dual 75-gallon fuel tanks. When the .50
gram background emission correction factor is added to the
component test average, the total test result is 1.28 g/test.
In its final written comments to this rulemaking, Ford included
full-SHED test results of a small (6,250 lbs. 7,000 lbs.
inertia weight) truck with a 4V carburetor, dual (19.5 and 19.0
gallons) fuel tanks, and Ford's California control system.
This vehicle's fuel system should generate significantly less
evaporative emissions than the fuel system tested for 1530
California certification because this vehicle's total fuel tank
capacity is only 26 percent of the total fuel tank capacity of
the fuel system tested for California. Yet, the full-SHED test
results averaged 122 percent (2.84 : 1.28 g/test) higher than
the component test results.
The above test procedure comparisons raise serious doubts
about claims that the Ford component test gives results th = -
closely parallel results from the full-SHED test. Other

-------
-27-
obvious flaws in the Ford component test procedure include: I)
no provision to evaluate the purging system, 2) no provision to
check storage capacities of vapor storage containers after they
have been loaded due to a hot soak, and 3) no provision to
check carburetor cola-soak losses or fuel tank hot-soak
losses. Ford's procedure would need much, development before it
would be an acceptable test for certification.
The component test procedure developed by GM is
significantly more sophisticated and, therefore, more expensive
than Fords. GM, like Ford, claimed that its component test
procedure gave evaporative emission loss numbers very close to
numbers generated by the full-SHED test procedure. GM tested
two HDGs by its component test procedure and then tested them
by the full-SHED test procedure. It should be noted that since
GM did not have a heavy-duty chassis dynamometer, it drove the
test vehicles over a road-route for preconditioning and
hot-soak warmup. GM did a total of 5 pairs of tests comparing
the component test to the full-SHED test. Three pairs were
done on vehicles with evaporative emission control system ana
two pairs were done on the same vehicles but in an uncontrolled
configuration. Since we are evaluating the two test procedures
in light of the fact that they would be used to test controlled
vehicles only, we will consider only those three pairs of tests
which used controlled vehicles.
The first problem with using GM's data is that GM was not
consistent in its manipulation of the intermediate test results
to arrive at a total test result. In two of the three pairs of
tests, GM compared the component test results to the full-SHED
test results and included the background emissions in the
fullSHED results. In the third pair of tests, GM subtracted
the background emissions from the full-SHED test total and then
compared that to the component test results which included
background emissions. GM gave no reason why background
emissions were not included in the full-SHED test results.
GM's figure for the full-SHED test was 3.88 g/test while its
figure for the component test was 3.98 g/test. This appears to
be good correlation. However, when the background emissions
are added to the full-SHED result to be consistent with the
other two pairs of tests, it becomes 6.38 g/test which is 60
percent higher than the component test result. We can
determine no reason why background emissions were not
included. Therefore, all three of the full-SHED test results
will include backg round emissions as do all three of the
component test results.
The next problem with comparing GM's component test
results to its £ull-SHED results is the measurement of
background emissions. GM claims that "background emissions" are
in large part aie to gasoline permeation through the fuel
lines. This source of HC emissions should properly be termed
as "evaporative" emissions instead of "background" emissions

-------
-28-
because it is an obvious part of the fuel system. The reason
that GM refers to fuel line permeation as a non-fuel or
"background* emission source is unknown. GM's component test
measures this permeation and then identifies the result as
"background" emissions. However, there is substantial data
which shows that paint, sealers, lubricants, tires and other
non-fuel components of automobiles and trucks give off HC
emissions. These sources are properly termed "background"
emissions.
In its final written comments Ford presented true
background emission data on a HDG. Ford removed the entire
fuel system, including the fuel lines, from the vehicle before
testing. This vehicle was in an incomplete configuration when
tested. That is, it was a chassis with engine and cab but no
bed or payload area. It would be likely that a payloaa area
with its significant amount of painted surface area would
result in somewhat higher background emissions than Ford's
results. The average of three tests showed diurnal background
emissions to be 0.09 g/test and hot-soak background emissions
to be 0.37 g/test for a total test result of 0.46 g/test.
In a joint testing program, EPA and Ford removed the
entire fuel system from a new 1977 HDG (and replaced it with a
propane fuel system). After 70 days, background emissions had
dropped to about 0.6 g/test. Furthermore, GM submitted data in
its final written comments which showed that three LDVs emitted
an average of 0.30 g/test of background emissions during the
hot-soak portion of the full-SHED test. These vehicles had
their fuel lines removed before testing so that none of the HC
emissions came from fuel line permeation. When we remember
that cold-soak (diurnal) background emissions would add to the
.30 g/test and that the LDVs are smaller than HDGS, GM's
results agree with Ford's as to the significance of true,
non-fuel emissions.
Since GM's full-SHED test for background emissions (in its
final written comments) includes true, non-fuel sources such as
paint, lubricants, tires sealers and sound deadeners in
addition to fuel line permeation, the component test for
background emissions should include these non-fuel sources as
well. Therefore, we have added 0.5 g/test to each of GM's
component test "background" emissions results to account for
the fact that GM's component test does not measure true,
non-fuel sources of evaporative HC while the full-SHED test
does.
The first vehicle GM tested was a 1979 heavy-duty pickup
truck with a 350 in3 engine, 4 bbl carburetor and dual, 20
gallon fuel tanks. GM tested this vehicle with its 1979
California evaporative control system installed and again, with
that control system plus a carbon element in the air cleaner.
The results of the two pairs of tests (component test procedure
vs. fnl1—SHED test orocedure) are as follows:

-------
-23-
Heavy-Duty Pickup with 1979 California Control System
Component Test
Full-SHED test
Component Test
Difference from
Full-SHED test
Diurnal
(g/test)
(background
not included)
.31
.34
-9%
Hot-soak
(g/test)
(background
not included)
.86
.67
+ 28%
Background Total
(g/test) (g/test)
1.24
.78
+ 59%
2.41
1.79
+ 35%
Heavy-Duty Pickup with 1979 Control System
Plus Carbon in Air Cleaner
Diurnal
(g/test)
(background
not included)
Component Test
Full-SHED Test
Component Test
Difference from
Full-SHED Test
The above
underestimated
overestimated
full-SHED test,
more than 35
.31
.35
-11%
results show
the diurnal
the hot-soak
The total
percent.
Hot-soak
(g/test)
(background
not included)
.43
.33
+ 30%
Background
(g/test)
1.24
.78
Total
(g/test)
1.98
1.46
+ 36%
and background
test results show
Certainly these
significant and the only conclusion that can
GM's component test procedure gives different
full-SHED procedure. If further testing
that GM's component test procedure
portion of the full-SHED test and
portions of the
a difference of
differences are
be drawn is that
results than the
showed that the
percent difference for each part cf the test procedure stayed
about the same for many different vehicles and configurations,
then GM's component test procedure would be usable since
full-SHED test results could be obtained from the component
test results by using correction factors. However, the results
from the other vehicle that GM tested show that the relative
differences between the two test procedures are unpredictable.
The second vehicle GM tested was a C-7 series tractor unit
with a 427 in3 engine, 4 bbl. carburetor, and dual, 50 gallon
fuel tanks. The results of this testing are shown below:

-------
-30-
C-7 Series Tractor
Diurnal	Hot-soak
(g/test)	(g/test)
(background	(background Background Total
not included)	not included) (g/test) (g/test
Component Test .88	1.83 1.77	4.48
Full-SHED Test .72	3.16 2.50	6.38
Component Test
Difference from
Full-SHED Test +22%	-42% -29% -30%
These results are completely opposite from the pickup results.
Where the pickup averaged .10 percent lower on the diurnal
portion of the test, the C-7 tractor is 22 percent higher.
Where the picku.p averaged 29 percent higher on the component
test hot-soak, the C-7 tractor is 42 percent lower. The
background and total test results were likewise opposite for
the two vehicles. We conclude that the GM component test
procedure will not predict full-SHED test results closer than
perhaps + 30-40 percent.
Besides the above test results, there are other,
qualitative, observations and questions that cast serious doubt
on GM's component test procedure. Listed below are four
problem areas that could cause significant differences in test
results between GM's component test and the full-SHED test.
These problem areas would have to be investigated further
before a component/mini-SHED test procedure would be acceptable.
Carbon Canister Equilibration GM suggests that carbon
canisters be equilibrated by flowing vapor into them at 4 scfh
for ten minutes and their purging at 50 scfh for twenty
minutes. GM does not present any evidence that the above flow
rates are typical of HD vehicles. The risk is that, although
the canisters may be equilibrated, they still could be
essentially "green". This question should be explored and
resolved.
Fuel Line Permeation GM states that vapor permeation
through fuel lines is a principal source of "background"
emissions. (It should be noted that EPA considers this source
to be part of the fuel system and, therefore, it is net
"background".) GM suggests that the hose be filled 75 percent
full of fuel and weighed for several days to determine the
permeation rate (grams/ft/hr) . The diurnal test result would
be the effective length of tubing for the vehicle times the
permeation rate. The hot-soak test result would be the
effective length times the permeation rate times 2.5. The "2.5
factor" is supposed to account for the increased temperature

-------
-31-
experienced by the fuel lines during the hot-soak. It is a
"rule cf thumb" factor for which GM provides no data but claims
"has been developed from experience." GM admits "hot bench
tests (on fuel lines) are relatively meaningless comparisons as
every hose experiences different temperatures depending on its
location in any specific vehicle. Temperatures can even vary
significantly from end-to-end of one hose." Thus, GM's "2.5
factor" for calculating hot-soak fuel line permeation from
coldsoak hose permeation should be investigated further to
determine the significance of temperature variation effects.
Then, the temperature of the fuel lines on each vehicle tested
should be measured and the correct hose permeation rate
applied. It is important to determine if the "2.5" factor is
accurate.
GM's suggestion to fill the hoses 75 percent full is
unsupported and needs investigation. The significance of hose
permeation is clear when one considers that GM determined that
.74 g/test to 1.27 g/test is due to this fuel system emissions
source.
Carburetor Hot-Soak Emissions. GM suggests that the
carburetor be mounted, with its mounting gasket, on an aluminum
block. The air cleaner and charcoal canister are also mounted
in their respective positions to the carburetor. An electric
heat source is applied to the aluminum block which heats the
carburetor fuel bowl. This heating is supposed to follow the
heating pattern on an actual vehicle.
The first problem that should be investigated further is
that of the convection currents which are set up due to the hot
engine block in the vehicle's engine compartment during a
full-SHED test vs. the hot aluminum block in GM's component
test. These air currents could be very different and could
influence the amount of vapor escaping through the air cleaner,
carburetor linkages, and carbon canister.
Another problem is that of the appropriate heating cycle.
Different vehicles will have different carburetor bowl heating
curves. Should one general curve be used or should each
evaporative family have its own?
A third problem concerns the background or non-fuel
emissions from the hot engine compartment. These emissions are
not measured in GM's component test. While it may be argued
that they should not be counted anyway, the standard would have
to be adjusted accordingly.
A fourth problem is the probability that there are sorr.e
evaporative emissions from the fuel tanks' during the hot-soak
test. The proximity of the fuel tank(s) to the exhaust system
may cause the tank(s) to absorb exhaust heat to the point where
the pressure of the vapor in the tank will be significantly
raised and evaporative emissions may occur.

-------
-32-
Diurnal Losses from Fuel Tank. The GM component/mini-SHED
procedure has no provision for the entrapment of heat around
the fuel tanks due to the payload/body platform. If a fuel
tank is so located as to be essentially surrounded (at least
the upper parts) by the vehicle's body, then heat from the
heating blankets can get trapped. This build-up of heat would
cause the vapor in the fuel tank to exert more pressure than if
the fuel tank is completely exposed to air circulation as it is
in GM's component test.
It is clear from the above discussion that actual emission
levels from HDGs would not be accurately known under a
component/ mini-SHED certification scenario unless a
substantial program to develop a reliable procedure was
undertaken. Air quality planners' confidence in the HDG
evaporative certification numbers to use for in-use emission
factors would be low. We will use the results of GM's
component tests to estimate the certification emission levels
for the component/mini-SHED certification alternative.
Although we don't have much confidence in GM's numbers as they
relate to the full-SHED test, they are the only numbers we
have. Assuming that manufacturers would take advantage of the
flexibility available in meeting the standard using this
procedure, we expect average emissions to be higher than with a
full-SHED test. Based upon the data presented above, we
estimate that the difference could easily reach 40 percent.
Thus, since the emission level under the Proposal alternative
was estimated to be 2.06 g/test, the certification emission
level under the component/ mini-SHED procedure is estimated to
be 1.4 times 2.06 g/test or 2.88 g/test. While this appears to
be below the sales-weighted standard of 3.47 g/test it should
be remembered that 2.88 g/test would be an average at
certification and, due to variability, it is probable that seme
vehicles would test above the standard.
The component/mini-SHED certification alternative would be
quite subjective since the different manufacturers would most
probably develop their own component test procedures. This
leads to problems of manufacturers equity and makes difficult
the implementation of manufacturer self-certification. EPA
could develop its own component test procedure but lack of
resources makes this possibility remote.
The Revised Proposal Certification Procedure
The major advantage to this certification alternative is
cost savings over the Proposal alternative. Because it is
anticipated that the amount of testing will be greatly reduced,
only one heavy-duty SHED should be needed per manufacturer.
For example, in its final written comments GM submitted its
estimate of its certification fleet. GM estimated that ur.cer
the proposed vehicle classification and selection scheme 19
emission-data vehicles would be tested for certification.

-------
-33-
These 29 vehicles have only four different carburetors among
them. Thus, if the emission control systems were the same for
all vehicles with the same carburetor, only four emission-data
vehicles would need to be built and tested. Even if we assume
two different emission control systems per family, then only
eight emission-data vehicles result. This still represents
more than a 70 percent reduction in the number of emission-data
vehicles. Thus, GM and Ford, who both claimed to need three
heavy-duty SHEDs under the Proposal should need only one under
this Revised Proposal alternative. We have determined that a
heavy-duty SHED will cost about $100K. Support equipment such
as heating blankets, emissions analyzers, temperature
achievers, etc. will add another $41K to the cost. Thus, the
industry total is expected to be 4 times $141K or S564K.
All of the four HDG manufacturers claimed they would need
at least one new heavy-duty chassis dynamometer. Ford claimed
it would need a total of three dynos (it has one already). GM
claimed it would need two dynos and IH and Chrysler each
claimed they would need one dyno apiece. Since this Revised
Proposal certification alternative will greatly reduce the
number of evaporative family-system combinations ana
consequently reduce certification, research and development
testing, the number of dynos required by each manufacturer will
be reduced. We had assumed that each manufacturer will
purchase one dyno. However, conversations with a dynamometer
manufacturer indicate that no new heavy-duty chassis dynos need
be purchased. The test procedure has been changed so that
light-duty dynos that have been converted to heavy-duty dynos
by adding inertia weights can be used for vehicle
pre-conditioning and warm-up. The dyno manufacturer has
indicated that a light-duty dyno can be upgraded to handle
13,500 lbs. inertia weight for about $25,000. Upgrading to
this inertia weight will handle all Class VI and below HDGs at
the new testing weight of 50 percent of GVW (see the "Test
Procedure" issue in this document). Thus, if we allow one
retrofit kit apiece for GM, Ford, IH, and Chrysler, then a
total of four kits would be needed. At $25K apiece, this comes
to $100K for dynos. Thus, total industry equipment costs would
be $664K.
Facility space costs would be substantially reduced
because of the reduction in the number of vehicles that need to
be built, soaked, and tested. GM claimed it would need 26,000
ft2 in new facility space. Our cost analysis indicates that
not only is 26 ,000 ft2 excessive for this revised proposed
alternative but it is a substantial overestimate for the
originally proposed alternative. We estimate that GM should
only require 4000 ft2 of facility space. We also estimate
that Ford would need 4000 ft2 while Chrysler and IH woula
each need 3000 ft2. This industry total of 14,000 ft-
times a building rental cost of $15/ft2-yr for 8 years gives
an industry total cf $l.7M.

-------
-34-
Th e only manufacturers to delineate certification testing
costs was GM. It claimed that $1,114 would be needed to build
the 29 emission-data vehicles required under the Proposal
certification alternative. This number of emission-data
vehicles should be reduced by at least two-thirds as a result
of the changes under this Revised Proposal certification
alternative. Thus, certification costs for GM should only be
about $350K. Allowing this same amount for Ford and half this
amount for IH and Chrysler would bring total industry
certification costs to $1.05 M.
Research and development costs are difficult to estimate
but should be quite small because the technology required is
well known. Under the Proposal and component/mini-SHED
certification alternatives we estimated $2M each for Ford and
GM and $1M each for IH and Chrysler which should be considered
liberal estimates. Under this Revised Proposal alternative,
R&D costs should be considerably less because of the reduction
in the number of evaporative family-system combinations which
would be tested for certification. In the Costs chapter of the
"Regulatory Support Document" we estimate that R&D will be
$2.0M for the industry.
Since control system hardware costs are assumed to be
equal under all of the certification alternatives, they do not
need to be estimated when comparing the differences in cost
among the alternatives. The total industry cost for equipment,
facility space, certification testing, and R&D under this
alternative is estimated to be $5.4M. The assumption that
control system hardware costs will be the same under all of the
five alternatives will be discussed further in the next section
where exceptions will be noted.
Another advantage of this alternative is that the level of
control would be known with a large degree of confidence. The
full-SHED test procedure will yield hard numbers which air
quality planners can use in estimating in-use emission factors
for air quality projections. We will assume that the average
certification emission level would be the same as that under
the Proposal certification procedure (i.e., 2.06 g/test). This
Revised Proposal alternative does not require as much
certification testing as the Proposal alternative because there
will not be as many family-system combinations. Therefore, one
might expect the average emission level to be closer to the
standard due to the reduced safety margin for variability that
the manufacturers have to include in their certification
emission levels. However, the Revised Proposal vehicle
.classification and selection criteria will result in
evaporative control systems which are somewhat overdesignea.
Manufacturers would install their worst case control systems on
some less than worst case vehicles so that they do not create
additional evaporative family-system combinations which would
have to be certified. Such overdesigned control systems will

-------
-35-
tend to control certification emissions to a level below that
which would occur under the Proposal alternative. Although we
cannot quantify these effects,, we do known they are relatively
small and work in opposite directions. We will assume that the
increase in the average certification emission level due to
less testing would be offset by the decrease in the average
certification emission level due to overdesignea evaporative
control systems on some vehicles. Thus, the emission level at
certification is estimated to be 2.06 g/test for this
alternative.
Another advantage to this alternative is that leadtime is
significantly reduced from that required under the Proposal
certification procedure. GM and Ford both claimed that the
major reason for extensive leadtime was the need to install
additional SHEDs and dynos. They claimed that lack of testing
equipment would severely slow down their R&D programs. Under
this Revised Proposal procedure the manufacturers would already
have the one SHED they need and the upgrading of light-duty
dynos to handle HDGs should take at most 6 months.
Furthermore, the amount of R&D would be substantially reduced
due to the reduction in evaporative emission' family-system
combinations.
Other advantages of this Revised Proposal certification
alternative include: it is an objective procedure to ensure
manufacturers' equity, it would allow easy implementation of
manufacturer self-certification, and the test procedure is well
known from light-duty experience.
A disadvantage of this alternative is that the problems of
incomplete vehicles, background emissions, SEAs and NCPs are
virtually the same as under the Proposal certification
procedure. For a general idea of these problems the reader is
referred back to the discussion of these disadvantages of the
Proposal alternative.
Another disadvantage to this Revised Proposal alternative
is that some KDGs would have evaporative emission control
systems that are overdesigned and, therefore, more expensive
than the minimal control system which would allow that vehicle
to just meet the standard. However, these extra hardware costs
are expected to be quite small and are offset by the decreased
cost of certification, research, and development testing. This
will be discussed further in the next section.
Abbreviated Certification. A major advantage of this
certification alternative is the great flexibility which it
would impart to the manufacturers.' The manufacturers would be
free to test as few or as many vehicles as they deemed
necessary to assure themselves that their HDG product lines
would meet the 3.0/4.0 g/test standard. Each manufacture:
could group its HDGs in any manner it chooses for development

-------
-36-
purposes. For example, a manufacturer might group together all
of his lighter-weight HDGs which are merely extensions of its
already controlled LDTs. The only difference affecting
evaporative emissions might be that HDGs have larger fuel tank
capacities. The manufacturer might only need to test the
"worst case" of these HDGs to assure itself that the other HDGs
in the group would also meet the standard.
Furthermore, the manufacturers would be free to use any
testing methods they want. In its comments to this rulemaking
GM claimed that its component test procedure correlated well
with the full-SHED test. Under this certification alternative
GM could use its component test procedure to develop its HDG
evaporative emission control systems. The manufacturers could
use full-SKED testing, component testing, engineering
evaluation and/or no testing in any mixture and amounts they
deem necessary in order to assure themselves that their HDGs
would meet the standard. The flexibility provided by this
certification alternative would allow the manufacturers to meet
the 3,0/4.0 g/test standard in the most cost effective manner
since all the decisions on needs for development and testing
expenses would be theirs.
Another advantage of this certification alternative is
that it would be the second least expensive of the five
alternative considered. Only the California-Type Engineering
Evaluation would be less expensive. Estimating the cost of
this Abbreviated Certification alternative is difficult because
it cannot be known how much testing each manufacturer will
consider necessary to assure itself of compliance. The cost of
this alternative will be somewhat less than the S5.4M
(excluding hardware) previously estimated for the Revised
Proposal alternative. Both alternatives include a 3.0/4.0
grams per full-SHED test standard, but this Abbreviated
Certification alternative does not require any testing.
Therefore, we will use the Revised Proposal cost estimates as a
starting point for estimating the costs of this Abbreviated
Certification alternative.
Under the Revised Proposal, we estimated that the
manufacturers would purchase 4 light-duty dynamometer retrofit
kits at a cost of $25K each. The amount of testing required
under this alternative should be less than under the Revised
Proposal, however it is not clear how much. Therefore, we will
allow the same number under this alternative. The same is true
for SHEDs and supporting test equipment. Thus, we have
estimated that total industry test equipment expenditures for
this alternative will be the same as for the revised proposal
alternative, that is, $66 4k. This is on the high side because
this alternative will allow the manufacturers to allocate their
resources more efficiently than the revised proposal
alternative to achieve the same result.

-------
-37-
The next area of costs to consider is that of new facility
space. Under the revised proposal alternative we estimated
that total industry costs for the facility space required by
this regulation would be $1.7M. This abbreviated alternative
would most likely allow some reduction in that amount because
the manufacturers would be free to use whatever testing method
they choose. Since the manufacturers claimed that a
component/mini-SHED test would give results comparable to the
full-SHED test, it is to be expected that much of the testing
will be done with this type of procedure. However, each
manufacturer will probably need at least some space for a
full-SHED test site and for parking vehicles. We have left the
amount of space the same under this alternative as under the
revised proposal alternative. Thus, the cost for facility
space under this abbreviated certification alternative is $1.7M.
Certification testing costs were estimated to be $l.05M
under the Revised Proposal alternative. Since the
manufacturers would only be required to submit a statement that
their vehicles would meet the 3.0/4.0 g/test standard if
tested, no certification testing would need to be done.
However, we presume that the manufacturers would do final
testing on their product lines before submitting their
statements of compliance. This cost has been estimated in the
"Costs" chapter of the "Regulatory Support Document." The
total industry cost for this "development testing" is projected
to be $8Q0K. This is about $205K less than the cost for
certification testing under the revised proposal alternative
because some family-systems will meet the standard so easily
that little or no final testing will be considered necessary.
R&D testing would still be necessary to assure each
manufacturer that his product line would meet the 3.0/4.0
g/test standard. We will assume that the amount estimated for
R&D under the Revised Proposal alternative would also be
required under this alternative. Thus, R&D costs are estimated
to be $2.QM for the industry.
The total industry cost for this Abbreviated Certification
alternative (excluding hardware) is estimated by summing the
estimates for new equipment, facility space, certification
testing and R&D. The sum of these estimates is $5.2M. This
compares to an estimate of $5.4M for the revised proposal
alternative with the reduction due to less certification
testing cost.
The level o£ control which would result from this
alternative is difficult to estimate since there is a lack of
experience with this type of certification procedure. We have
assumed that the manufacturers would do about the same amount
of testing under this' alternative as they would under the
Revised Proposal alternative. If the amount and type of RSD
testing is the same for both alternatives then the level Cl
control would be expected to be about the same. However, the

-------
-38-
manufacturers would be free to use component/mini-SHED testing,
bench testing or no testing as they deem necessary. The use of
these testing methods (as compared to full-SHED testing) will
tend to increase the level of emissions to somewhere above that
for the Revised Proposal alternative. Since we do not have a
firm estimate of the level of control for this Abbreviated
Certification alternative, we will express the expected level
of control as a range instead of a single number. The upper
bound of the range has been chosen as the midpoint between the
level of control expected with the Revised Proposal alternative
(2.06 g/test) and the sales-weighted standard (3.47 g/test).
This midpoint is 2.77 g/test. The lower bound has been chosen
as the Revised Proposal level of control (2.06 g/test).
As discussed previously, another important factor to
consider in selecting a certification procedure is its
objectivity. in general, objectivity is necessary to assure
that manufacturers are treated equally and to make
certification as routine as possible. Manufacturers' equity
would not be a concern during certification since all
manufacturers would be assured certification if they submit the
required statement of compliance (and any other information
which EPA may request), and act in good faith. However,
manufacturer's equity should be considered for the case of
later enforcement options. This alternative is objective in
that respect because all manufacturers will be subject to the
full-SHED test procedure for later enforcement and that test
procedure has been developed over a number of years for
repeatability and for minimization of subjectivity. Therefore,
this certification alternative will be considered as one of the
more objective ones.
Another advantage of this alternative is that it would
allow plenty of leadtime for the manufacturers to design and
develop their evaporative emission control systems before the
implementation date of the regulation. Required leadtime wculo
be substantially reduced from the Proposal alternative because
manufacturers would neeo to do little if any facility
modification before launching their R&D programs. As discussed
in the issue titled "Leadtime" (in this document), the leactime
between publication of the final regulation and the date of
implementation shculo be more than adequate for the
manufacturers to aevelop their control systems.
The biggest disadvantage with this alternative is that the
level of control, although estimated to be fair (2.77 g/test)
to excellent (2.06 g/test), would not be known with n.ucr.
confidence. It is conceivable that a manufacturer(s) cclIc
decide to reduce its effort during development of its control
systems from that level of effort needed to assure that all of
its HDGs meet the stancaro. However, EPA believes that cr.t

-------
-39-
risk of noncompliance is small given that the standard can
easily be met/ and that EPA retains the right to do
confirmatory testing.
This Abbreviated Certification alternative would clearly
depend on the manufacturers' honesty and integrity. Our basic
approach to this alternative is one of relying on the
manufacturers to exert a good faith effort. EPA retains
authority to do confirmatory and/or in-use testing for
compliance, but we do not have plans at this time to exercise
that authority on a routine basis.
The problem of background emissions would be substantially
reduced under this alternative. Under the Proposal alternative
manufacturers would have to age their emission-data vehicles to
stabilize background emissions before certification testing.
Under this Abbreviated Certification alternative there would be
no certification testing, and therefore, no problem. For R&D
purposes, manufacturers can use any testing method(s) they
choose. Since the manufacturers claimed that a
component/mini-SHED test gives results which closely correlate
with the full—SHED test procedure and background emissions are
not a problem when a component/mini-SHED test procedure is
used, background emissions during R&D work should be a minimal
problem.
If SEA testing of HDGs ever occurred, then the problem
discussed under the Proposal alternative would also exist under
this alternative because the SEA would test vehicles by the
ful1-SHED method.
The incomplete vehicle problem is little changed under
this alternative as compared to the Proposal alternative. This
Abbreviated Certification procedure retains the SHED test
procedure and, therefore, if secondary manufacturers were
required to show compliance because of modifications and/or
additions, then the. method (full-SHED testing) of proving
compliance would be very burdensome for those secondary
manufacturers. The reader is referred to the issue title
"Incomplete Vehicles" for our analysis of and recommendation
for this problem.
Comparison of the Alternatives
This section will compare the five certification
alternatives. Table 2 summarizes the advantages and
disadvantages of the five alternatives. It is obvious that the
Proposal certification alternative can be withdrawn frcm
further consideration. The Revised Proposal alternative is as
good as or better than the Proposal alternative in all but cne
of the categories. While that one exception (i.e., the
confidence in the level of control at certification for
application to in-use emissions estimations) is important, the

-------
Table 2
Comparison of the Five Certification Alternatives
Level of Control
at Certification
(gpt)
Proposal
best
(2.06)
Confidence in Level
of Control (in-use
emissions estimates)
Cost Excluding Hard
ware ($M)
Objectivity (mfr.
equity and self-
cert. )
best
worst
(25)
best
Engineering Component/ Revised
Evaluat ion Mini-SHED Proposal
worst
(4 .09)
worst
best
(0.5)
worst
fair
(2.88)
poor
fair
(11)
poor
best
(2.06)
good
good
(5.4)
best
Abbreviated
Cert if ication
fair
(2.06-2.77)
fair
good
(5.2)
good
Leadtime	worst
Incomplete Vehicle worst
Problem
best
best
fair
good
fair
poor
good
poor
Background
Emissions, SEA and
NCPs Problem
worst
poor
best
poor
poor

-------
-41-
difference between the Proposal arid Revised Proposal in this
category is not very great and the superiority of the Revised
Proposal in- other categories more than compensates for the
deficiency in this one category.
Table 2 also shows that the Revised Proposal procedure is
as good as or better than the component/mini-SHED procedure in
all but two categories. Both of these two categories ace
relatively unimportant compared to the other categories. The
need for and desirability of SEAs and NCPs for evaporative
emission control of HDGs is minimal. While background
emissions would be less of a problem under the
component/mini-SHED alternative, we have determined that this
problem is not a large one and it has workable solutions. The
discussion on background emissions found in the issue
"Technological Feasibility" presents analyses and solutions to
the problem.
The incomplete vehicle problem could be considerably
diminished under the component/mini-SHED certification
alternative. But since the Revised Proposal offers so many
more advantages, it is one problem that we would accept. The
anaIysis and solution of this problem can be found in the issue
entitled "Incomplete Vehicles."
Having withdrawn two of the five alternatives from
consideration because of the obvious and overwhelming
advantages of the Revised Proposal procedure, we are left with
the Engineering Evaluation and the Abbreviated Certification
alternatives to compare with the Revised Proposal procedure.
Since it is clear that the Revised Proposal will control
emissions to a lower level than the other two alternatives, the
next question is: "Is the extra control worth th§ extra cost?"
In order to answer that question we have calculated the
marginal cost-effectiveness of moving from the Engineering
Evaluation alternative to the Revised Proposal alternative. We
have also calculated the marginal cost-effectiveness of moving
from- the Engineering Evaluation alternative to the Abbreviated
Certification alternative and of moving from the Abbreviated
Certification alternative to the Revised Proposal alternative
as will be discussed later. The difference (4.09 g/test to
2.06 g/test) between the control level of the Engineering
Evaluation ana the Revised Proposal is 2.03 g/test. The cost
difference between these alternatives is $4.9M. The marginal
cost-effectiveness number will be the increase in cost per
gasoline-fueled heavy-duty vehicle as one moves from the
Engineering Evaluation to the Revised Proposal divided by the
decrease in lifetime HC emissions per HiG (in tons]. The
methodology for converting g/test tc grams per mile Igor.) is
given in Chapter IV of the "Regulatory Support Doc urn en t" to
this rulemaking. Using that methodology we have calcciaf.ec
that the difference between the two alternatives jrcet

-------
-42-
consideration is .179 gpm. Since the typical lifetime for a
HDG is 114,000 miles, the difference between the total tons of
HC emitted over the lifetime of an HDG for the two alternatives
is;
(.179 gpm) x (114,000 mi) x (45^^bg) x (2ooo°lbs) = -0255 tonS
The difference in cost between the two alternatives is
$4.9M. This cost difference must be discounted to 1984 and
then amortized over 5 years production to arrive at a per
vehicle cost. Discounting $4.9M for 3 years to 1984 (@ 10
percent) yields $6.5M. Total HDG production for 1984 through
1983 inclusive is estimated to be 1,768,300 (see Chapter III of
the "Regulatory Support Document") or an average annual
production of 353,780 HDGs. Amortizing $6.5M over 5 years at
353,780 HDGs per year yields a value of $4.85 per vehicle.
Dividing the difference in price by the difference in tons of
HC emitted per vehicle gives a cost-effectiveness of $216/ton
HC.
IX.
This marginal cost-effectiveness number is well within the
range of other HC control strategies. Table VI-A in Chapter VI
of the "Regulatory Support Document11 shows various HC control
strategies and their cost-effectiveness numbers. It should be
emphasised that the number we have calculated above is the
marginal cost-effectiveness of going to the Revised Proposal
certification alternative from the Engineering Evaluation
alternative. In general, the marginal cost-effectiveness of
controlling the last several percent of an HC emission source
increases geometrically as the one-hundredth percentile is
approached. The table of cost-effectiveness in Chapter VI of
the "Regulatory Support Document" shows this tendency. The
cost-effectiveness of controlling the first 50 percent of an HC
source is much lower than the marginal cost effectiveness of
controlling the last 10 percent. .-j3S-;;:::5^trt^of this fact, a
marginal cost-effectiveness of just^$216/ton at such a high
level of control is, indeed, a very__5M^d-^rnlmber. Therefore,
the extra HC control achieved by the Revised Proposal
alternative is we11 worth the extra expense.
The Revised Proposal alternative is also better than the
Engineering Evaluation alternative in the categories of
"Confidence in Level of Control" and "Objectivity." These are
important considerations in selecting a certification
procedure. The Revised Proposal alternative will give hare,
full-SHED test emission numbers during certification. These
numbers will be relatively easy to use when air quality
planners need to project the level of in-use HDG emissions.
The Engineering Evaluation will give no emission numbers at all
and, over tijr.e, the actual emission levels of HDGs would be
expected to vary greatly. The Revised Proposal alternative is
objective and will assure manufacturers' equity at
certification. The Engineering Evaluation alternative would
resu.lt in many different types of evaluations which could leac

-------
to inequitable treatment of one manufacturer's submittal as
compared to another's.
Required leadtime is less under the Engineering
Evaluation. However, there would be sufficient leadtime to
implement the Revised Proposal for the 1984 model year. Since
there is enough leadtime for both alternatives, this apparent
advantage of the Engineering Evaluation is not an important one.
The problem of incomplete vehicles and the problem of
background emissions would disappear with the Engineering
Evaluation alternative. However, the advantages, of the Revised
Proposal alternative would outweigh these disadvantages. We
would accept these problems and develop workable solutions to
them. Finally, SEAs and NCPs are not possible at all under the
Engineering Evaluation while they are possible under the
Revised Proposal alternative.
Therefore, we have concluded that the Revised Proposal
would be preferable to the Engineering Evaluation. So far the
Revised Proposal alternative has been judged superior to the
Proposal alternative, the compcnent/mini-SHED alternative and
the Engineering Evaluation alternative. This leaves only the
Abbreviated Certification alternative to consider.
We have also calculated the marginal cost-effectiveness of
going from the Engineering Evaluation to the Abbreviated
Certification and of going from Abbreviated Certification to
the Revised Proposal. As discussed above, the marginal
cost-effectiveness of going from Engineering Evaluation with a
control level of 4.09 g/test all the way to the Revised
Proposal with a control level of 2.06 g/test is estimated to be
$216 per ton HC. The estimated level of control under the
Abbreviated Certification alternative (2.77 g/test) lies
between these other two alternatives. If we assume that the
Abbreviated Certification level of control would be 2.06
g/test, then the marginal cost-effectiveness of going from
Engineering Evaluation to Abbreviated Certification is $207 per
ton HC. If we assume that the Abbreviated Certification level
of control would 2.77 g/test then the marginal
cost-effectiveness would be $318 per tone HC. Thus, the
cost-effectiveness of moving from the Engineering Evaluation
alternative to the Abbreviated Certification alternative is
very good regardless of where, within the range assumed, the
actual Abbreviated Certification level of control turns out to
be.
The marginal cost-effectiveness of going the rest of the
way to the Revised Proposal may or may not be good depending on
the level of control assumed for the Abbreviated Certification
alternative. It we assume a level of control of 2.77 g/test,
then the marginal cost-effectiveness of moving from Abbreviated
Certification to Revised Proposal is only $25 pet ton HC.

-------
-44-
Howevec, if we assume an Abbreviated Certification level of
control of 2.06 g/test, then the marginal cost-effectiveness
becon.es infinite becaue costs are increasing but benefits
regain the same.
The above calculations show that it is cost effective to
move from Engineering Evaluation to Abbreviated Certification.
However, whether or not it is cost effective to go the rest of
the way to the Revised Proposal depends on the actual level cf
control obtained under the Abbreviated Certification
alternative. We believe, as indicated by our inclusion of
substantial r&d effort when estimating the cost cf the
Abbreviated Certification alternative, that the manufacturers
will make a good faith effort to control evaporative
emissions. Although we do not normally expect to require the
manufacturers to submit their data and/or analyses and we do
not routinely expect to do confirmatory testing, we believe
that the risk of EPA exercising these options is sufficient to
assure compliance with the appropriate standard. For this
reason, we conclude that the level of control obtained unoer
the Abbreviated Certification alternative will be closer to the
lower end (2.06 g/test) of the range ana the cost effectiveness
of moving to the Revised Proposal is likely to be very poor.
Furthermore, the Abbreviated Certification alternative allows
the manufacturers substantial flexibility in the total effort
expended as well as the division cf that effort over their
product lines and the scheduling of that effort. This
flexibility increases the probability of maximum efficiency cf
resources to obtain the same control level. Thus, the cost
differential between the Abbreviated Certification alternative
and the Revised Proposal alternative is probably greater than
cur conservative estimate.
For all the above reasons and in order to lessen the
burden on this industry to the extent possible we recommeno
that the Abbreviated Certification alternative be included in
the Final Rule.
Recommendation
We recommend that the Abbreviated Certification
alternative Le adopted as the certification procedure tor
control of evaporative emissions from HDGs. This procecure
should result in control of emissions to a level Lelcw the
3.0/4.0 g/test stanoaro at 'a low cost. It will give the
manufacturers a great ceal of flexibility in their approaches
to the oevelopment ario testing cf their control systems. t-.cst
of the emission reduction expected from the original proposal
will be obtained at a much lower overall cost.

-------
-45-
B. Issue; Incomplete Vehicles
Summary of Issue
The four HDG manufacturers (GM, Ford, Chrysler, and
International Harvester) sell many vehicles in an incomplete
form. Most of these incomplete vehicles are Class IV and above
(greater than 14,000 lbs. GVW) heavy-duty vehicles. These
vehicles may be complete except for the payioad bed or box or
the vehicle may only include an engine and a chassis. The
purchaser (secondary manufacturer) of these incomplete vehicles
completes them by adding engine compartments, operator
enclosures, payioad devices and/or fuel tanks and lines.
Secondary manufacturers also modify already complete
vehicles by adding extra fuel tanks or changing the location of
existing fuel tanks and/or exhaust systems. All of the above
manipulations of HDGs by secondary manufacturers could affect
the level of evaporative emissions from those vehicles. Herein
lies a problem faced by EPA in its attempt to promulgate this
HDG evaporative emissions regulation. Although completed
vehicles meet the HDG evaporative emission standard because
they are tested during the certification process; how can we
make sure that vehicles which are later added to or modified by
secondary manufacturers also meet the HDG evaporative emission
standard?
The most obvious answer would be to require secondary
manufacturers who make additions or changes to vehicle
parameters that could reasonably be expected to affect the
level of evaporative emissions to test those vehicles for
compliance. However, there are hundreds of secondary
manufacturers many of which are quite small in terms of total
assets. The expense of a heavy-duty SHED, its associated
analysing bench, and space to put it along with the expense of
a heavy-duty dynamometer would probably eat up many years of
profit for many of the smaller secondary manufacturers. Thus
f ull-SHED testing by secondary manufacturers would be very
burdensome and would force many out of business.
In the NPRM the Agency proposed that the -primary
manufacturers would determine the limits of a worst case
completed vehicle configuration for each incomplete vehicle
they wish to market. These limits (in addition to those
specified by the exhaust emission certificate) would be those
defined by the evaporative emission family-system combination
and the evaporative vehicle configuration, i.e., fuel tank
volume, carburetor bowl fuel volume, method of vapor storage,
vapor storage material, vapor storage working capacity, i.idt::od
of carburetor bowl venting, vapor purging technique, fuel
system, maximum GVWR, maximum frontal area, boay type, and
other features as specified by the Administrator. If an
incomplete vehicle was selected to be an emission-data vehicle,

-------
-46-
th en the manufacturer would build up that incomple vehicle in
accordance with the worst case limits he has previously
determined. The incomplete vehicle would then be covered by
the certificate of conformity as long as the secondary
manufacturer stayed within the worst case limits. If the
secondary manufacturer wanted to make an addition or
modification which was worse than the primary manufacturers
worst case limits, then the secondary manufacturer would have
to test the vehicle and show compliance.
Summary of Comments
A total of eight organizations commented on this issue of
incomplete vehicles. Besides the four HDG primary
manufacturers (GM, Ford, Chrysler, International Harvester),
four trade associations submitted comments. These were the
Motor Vehicle Manufacturer Association (MVMA), the National
Truck Equipment Association (NTEA), the Truck Body and
Equipment Association (TBEA), and the National Automobile
Dealers Association (NADA).
The commenters generally claimed that EPA failed to define
"worst case" adequately. As discussed above, EPA listed a
number of parameters for which the manufacturer would neeu to
determine worst case limits when certifying an incomplete
vehicle. GM stated that in some cases the completed vehicle
with the largest frontal area may not be the completed vehicle
with the highest GVW. If such a chassis/engine combination was
picked for emission-data testing, then the resulting
emission-data vehicle build would be worse (in regards to
evaporative emissions) than any vehicle that would actually be
produced. This was considered unfair.
The commenters also claimed that a worst case
determination for some of the parameters on EPA's list would oe
impossible because EPA had failed to define how changes in such
parameters affect evaporative emissions. For example, the
proposal would require the manufacturer to determine a worst
case body type but EPA failed to delineate how a manufacturer
could determine which possible body type was worst. The
inclusion of the undefined and open-ended worst case parameter
of "other features as specified by the Administrator" was
claimed to be unreasonably vague. Since commenters considered
the proposed scheme for determining worst case vehicles to be
unreasonable, impracticable, and not objective, it was clairaeu
to be illegal as shown in Paccar, Inc. vs. NHTSA, 573 F. 2d 632
(9th Cir. 1973); Chrysler vs. Dot, etal, 472 F. 2d 659 (5c h
Cir. 1972 ); I.H. ezal vs. Ruckelshaus, 486 F. 2d. 375 (D.C.
Cir. 1973).
Another major concern with EPA's proposed solution to the
incomplete vehicle problem was the claim that even if it were
known how to deter:..ine the worst case of a parameter, finding

-------
-47-
the completed vehicle which incorporated that worst case would
be impossible. GM. stated that "a large portion of these
(incomplete), vehicles are ordered through dealerships; for
these order, General Motors has no direct knowledge of who the
purchaser is, much less of what kind of finished body or cargo
enclosure the secondary manufacturer intends to construct."
Furthermore, GM claimed that only in a limited number of sales
does it have contact with a secondary manufacturer and "even
for these orders we frequently have little or no knowledge of
the final form, the method of construction, and the materials
used on the completed vehicle. Also, "the secondary
manufacturer may not know the final form of the vehicle when he
contracts with General Motors or its dealer to provide a
chassis. It is common practice for such manufacturers to order
chassis in speculation of subsequent resale and then finish the
vehicle construction on receipt of a final customer order."
Thus, GM states that "the blatant unreasonableness of requiring
a manufacturer to 'seek and find1 worst cases in the market
place alone should have prevented this regulation from being
proposed..."
Ford, Chrysler and IH agree that due to the unlimited
variety of secondary manufacturer modifications and the fact
that these modifications are done by hundreds of secondary
manufacturers outside the primary manufacturers' knowledge,
EPA's proposed scheme is unreasonable and overly burdensome.
IH included a Federal Register listing of several hundred
bonafide motor vehicle manufacturers. GM and Chrysler stated
that even if the above problems were to be resolved, the
prevention of sale of an engine/chassis/fuel system combination
in any form because a worst case configuration fails the
certification test is obviously unjust.
Another problem that the primary manufacturers expressed
concern about was that they would lose control of their
certification timing. They claimed that once they produced the
engine/chassis for an emission-data vehicle they would have to
ship it to a secondary manufacturer in order to complete it in
its worst case configuration. They claimed that this could
take months and they wouldn't have any control over the work.
This would delay their certification program so much that tney
might not be able to finish certification before production was
scheduled to begin.
Beyond the practical problems of defining and t.ien
discovering a worst case final configuration, the commencec3
claimed that the proposal would impose vicarious liability on
the primary manufacturer. They claimed that seconJucy
manufacturers perform many additions and/or modifications -~.ii.ch
could influence evaporative emissions. Some of these addi:;jn3
or modifications have - not been accounted for in .i^A's
proposal. For example, relocations or revisions to exhaust
systems, the use of paints, sealers and sound deadeners, trie

-------
-48-
installation of permeable plastic fuel tanks, and the
particular placement of operator's enclosures in relation to
exhaust system components can all affect evaporative emissions
but the proposal does not address these items. Thus, the
primary manufacturer could certify an incomplete vehicle which
is then completed by a secondary manufacturer who stays within
the worst case limits set by the primary manufacturer. The
completed vehicle would be covered by a certificate of
conformity. However, the vehicle could later be found to be in
noncompliance because of additions and/or modifications for
which there were no worst case limits. The commenters claimed
that the proposed regulation would make the primary
manufacturer liable even though noncompliance was not their
fault.
GM and MVMA claimed that Congress clearly intended that
the incomplete vehicle manufacturer should warrant the vehicle
it produces and the subsequent manufacturer should warrant the
product it produces (i.e., the remainder of the vehicle as
completed by it). They cited Section 216(1) of the Clean Air
Act which defines "manufacturer" as any person engaged in the
manufacturing or assembly of new motor vehicles or new motor
vehicle engines. GM cited four cases in which the court found
that the Administrator lacks power to impose liability upon a
blameless party for the acts of another beyond his control.
These cases include: Chrysler Corp., etal vs. EPA, 600 Fed 2d
904 (1979); Amoco Oil Co. vs. EPA, 177 U.S. App. C.C. 123, 543
F. 2d. 270 (1976); Amoco Oil Co. vs. EPA, 163 U.S. App. D.C.
162, 188-180, 504 F. 2d. 722, 748 (1974); and Rex Chaimbelt,
Inc. vs Volpe, 486 F. 2d. 757, 762 (7th Cir. 1973). The other
primary manufacturers also claimed that the proposal imposed
vicarious liability on them and, therefore, it was illegal.
The primary manufacturers were in general agreement on a
recommended solution to the incomplete vehicle problem. They
stated that EPA should only require that the vehicle meet
emission control requirements at the time that it left the
primary manufacturers control. Ford stated that incomplete
vehicles should be certified in their "incomplete" state. GM
suggested that EPA could handle the liability problem in a
similar way to how the Agency handled the noise control
regulations liability problem.[1] The procedures developed
from that case permit the manufacturer to state that the
incomplete vehicle complied with applicable EPA noise control
regulations at the time it left the manufacturer's control.
GM included labeling language which would be placed on all
incomplete vehicles leaving the primary manufacturer's
control. GM suggested the following label;
"This vehicle conformed to the U.S. EPA certification
regulations applicable to 19	 Model Year New Motor Vehicles at
the time it left (insert initial incomplete vehicle
manufacturer's name) control, provided it was not equipped with
a temporary fuel system.

-------
-49-
HOTICE
Any subsequent manufacturer who installs any permanent fuel
system or modifies an existing permanent fuel system on this
vehicle is required to obtain prior approval from the
Environmental Protection Agency."
GM goes on to say that the above label would require EPA
to develop a clear testing procedure for certification of
incomplete vehicles and to develop a means by which secondary
manufacturers could obtain approval for new, permanent fuel
systems or modifications of existing systems.
MV MA suggests that "EPA coulu require the primary
manufacturers to certify that an incomplete vehicle is capable
of being made to conform to applicable emissions regulations
when completed within the manufacturer's fuel system
specifications. Such a statement could accurately reflect the
extent to which an incomplete vehicle manufacturer can predict
compliance of a vehicle whose eventual design and end-use are
unknown at the time when the statement is made."
The three trade associations (NTEA, NADA, and TBEA) that
represent secondary manufacturers were concerned that the
primary manufacturers would not include a wide enough range for
the worst case limits. NTEA stated that the primary
manufacturer does not know what the worst case limits should be
because of insufficient information. These trade associations
all commented that most secondary manufacturers do not have the
resources to purchase the necessary test equipment to run a
full—SHED test. Thus, many secondary manufacturers might be
forced out of business thereby leaving the business to "large
secondary manufacturers and 'shadetree1 mechanics." TBEA
stated that primary manufacturers don't offer enough options
and modifications for all truck customers and, therefore, a
buyer will no longer be able to purchase the truck that really
meets his requirements. NADA pointed out that the price of
HDGs would be raised due to any testing costs. NADA also
suggested that EPA expand "worst case" to include all
modifications by secondary manufacturers and require the
primary manufacturers to inform the secondary manufacturers of
such worst case limits.
Analysis of Comments
The comments received on this complex issue were very
useful and have helped to clarify the problems involved. in
summary, the primary manufacturers had three main areas of
concern. First, they felt that some of the parameters listed
in the NPRM were related to evaporative emissions in a vague
and ill-defined way so that determining a worst case situation
would be very diciiLcult. Second, they felt that the burden of
finding all of the modifications and/or additions to incomplete

-------
-50-
vehicles in order to determine a worst case vehicle was
intolerable. Finally, they were concerned about their
liability if in-use testing showed a modified vehicle to be in
noncompliance. This analysis will address the first two areas
in a combined way and then the liability question will be
discussed.
Information obtained from the four primary manufacturers
(see the docket) indicate that the number of HDGs which leave
the factories in an incomplete form is substantial. Ford
indicated that 55 percent of their medium-duty gasoline-fueled
trucks (8500-11,000 lbs. GVW) are incomplete and 100 percent of
their 16,000 lbs. and above GVW vehicles are sold as
incomplete. (Ford produces very few HDGs in the 11,u00-16,000
lbs GVW range.) Chrysler told us that approximately 10 percent
of their 8500-10,000 lbs. GWI HDGs are incomplete and virtually
all of their 10,000-14,0 00 lbs. GVW HDGs are incomplete.
(Chrysler doesn't produce HDGs over 14,000 lbs. GVW). IH only
produces HDGs over 16,000 lbs. GVW and it indicated that
virtually all of such vehicles leave its plants in an
incomplete form. GM indicated that about 30 percent of its
HDGs below 14 ,000 lb. GVW and about 80 percent of those above
16,000 lbs. GVW are sold as incomplete vehicles. Incomplete
fuel systems and" incomplete cargo areas are the major items
that are not finished. The total number of incomplete HDGs
approaches 50 percent of all HDGs produced. This is
considerably more than we had believed when this regulation was
proposed.
In addition there are hundreds of secondary
manufacturers. The list of bonafide motor vehicle
manufacturers that IH submitted has 256 companies on it. This
list was published in the Federal Register on June 23, 1980.
It contains only those companies that wish to obtain duty-free
Canadian articles. Thus, it is likely there are a number o£
secondary manufacturers which are not on the list because they
do not import Canadian articles. Also, there are probably some
secondary manufacturers who could be on the list but have
chosen not to be or are not aware of the list. perhaps a more
accurate assessment of the number of secondary manufacturers is
the membership total of the Truck Body and Equipment
Association. TBEA currently has 700 members.
When the large number of incomplete vehicles produced is
combined with the fact that there are hundreds of secondary
manufacturers, the problems involved in expecting the primary
manufacturers to determine the worst case for each evaporative
emission parameter listed in the NPRL4 become obvious. ir.e
primary manufacturer would have to contact every customer of
incomplete HDGs it has and attempt to determine what additions
and/or modifications to each listed parameter the customer
plans to make. Some of those customers would ' be dealers vino
buy incomplete vehicles without knowing to whom they wL1i

-------
-51-
ultimately be sold, let alone what kind of modifications and/or
additions will eventually be made. These dealers typically buy
chassis/cab combinations on the expectation that someone will
want to purchase them at some future time. The dealer has no
knowledge of who the ultimate purchaser might be or what the
final, completed vehicle will be like.
Not only would the primary manufacturers have to
investigate in detail each of its customers' vehicle completion
plans but for some of the evaporative emission parameters
listed in the NPRM a worst case determination would be quite
difficult. For example, little is known about the effect of
"body type" on evaporative emissions. While different body
types undoubtedly trap heat around the fuel tank(s) and line(s)
to different degrees thus changing the amount of evaporative
emissions, there is no data on the extent that a certain body
type might increase emissions over one or more other body
types. In the 1JPRM, we assLiaed that there were relatively few
incomplete vehicles sold and, therefore, the effort to
determine which body types were worst case would be minor.
However, in light of the fact that there are such a ^reat
number of incomplete vehicles sold, most of which are sold
without the body, the burden placed on the primary
manufacturers would be excessive. Therefore, "body type" is
omitted from the final rule.
EPA also included, in the NPRM, an open-ended statement of
evaporative emission parameters. This statement was "and other
features as specified by the Administrator" and was included so
that we could add additional parameters if we discovered the
need to do so. Because we believed the number of incomplete
vehicles was relatively small, we considered the potential
burden of such an open-ended statement to be minimal. However,
since the number of incomplete vehicles is substantially
greater than we had expected, we agree with the manufacturers
who claimed that such an open-ended requirement is unreasonably
burdensome and we are recommending that the statement be
dropped.
Another parameter included on the list in the NPRM was
flfuel system™ . This term refers to the fuel lines, their
routings, the fuel tank, the fuel pump and the carburetor.
(The NPRM listing includes fuel tank volume and carburetor bowl
fuel volume as separate parameters.) While it is doubtful that
secondary manufacturers would modify the fuel tank (unless they
are adding extra fuel tank volume), the carburetor or the fuel
pump, they do occasionally reroute and modify fuel lines in
order to accommodate the body type. The location of a fuel
line with respect to heat sources can affect evaporative
emissions since fuel lines permeate HC vapors in proportion to
temperature exposure. Furthermore, different fuel line
materials will permeate HC vapors at different rates. Under
the incomplete vehicle requirements in the MPRM the primary

-------
-52-
manufacturers would have had to determine which modifications
to the fuel lines would be worst case. Again, due to the large
number of Incomplete vehicles, the proliferation of secondary
manufacturers/ and the lack of hard data on the extent that
increased temperatures would increase HC vapor permeation we
are recommending that fuel lines be deleted from the incomplete
vehicle requirements except that secondary manufacturers will
have to use non-metal fuel line material which is at least as
impermeable as the material used by the primary manufacturer.
The omission in the final rule of "body type," "fuel
lines" and "other features as specified by the Administrator"
should substantially decrease the burden on the primary
manufacturers. While these parameters affect evaporative
emissions to some extent, we believe the effect is not large
and that the increase in emissions due to these factors will be
small. We are also recommending that the parameters of maximum
GVWR and maximum frontal area be dropped from the list. It
would be difficult to determine a maximum frontal area because
of the custom orders received by secondary manufacturers.
Furthermore, this parameter should not have a major impact on
evaporative emissions anyway. In regards to maximum GVWR, the
primary manufacturer sells a chassis/engine combination with
some designated maximum GVWR limit based on the strength of the
chassis. This is necessary so that the secondary manufacturer
does not complete the vehicle and then recommend a maximum GVWR
which is too large for the strength of the chassis. Therefore,
the manufacturers should know the proper GVWR of each
incomplete vehicle for evaporative emission control system
development purposes and that control system should be
sufficient (at least in terms of GVWR) for the completed
incomplete vehicles. In summary then, we recommend that "body
type", "fuel system", "maximum GVWR", "maximum frontal area"
and "other features as specified by the Administrator" be
deleted from the list of evaporative emission parameters to be
considered in the treatment of incomplete vehicles because even
their combined effect on evaporative emissions is expected to
be minimal and the burden on the primary manufacturers if they
were required to find each worst case could be excessive.
The other parameters listed in the HPRM should be retained
in the final rule. These parameters are carburetor fuel bowl
volume, method of vapor storage, vapor storage material, vapor
storage working capacity, method of carburetor bowl venting,
and vapor purging technique. (Fuel tank volume will be
discussed separately below.) These six parameters of the
evaporative emission control system should not and probably
would not be Modified by secondary manufacturers. When a
secondary manufacturer completes an incomplete vehicle by
adding the body or adding an operator's enclosure there should
be no reason why that secondary manufacturer needs to modify
and/or add to these, six control system parameters. For
example, a secondary manufacturer shouldn't need to switch

-------
-53-
carburetors, which might change the fuel bowl volume and the
method of carburetor bowl venting. The engine that was ordered
with the chassis should have the desired carburetor. The vapor
purging technique will likely be engine specific also and,
therefore, the secondary manufacturer should not need to modify
it. Certainly the secondary manufacturer should have no reason
to change the method of vapor storage or the vapor storage
material. The vapor storage working capacity might need to be
modified if the secondary manufacturer added fuel tank capacity
above that for which the primary manufacturer had designed the
control system. This potential problem will be discussed below
under the topic of liability.
These six parameters, in addition to being listed in the
NPRM for incomplete vehicle consideration, are evaporative
control family-system determinants in the proposed as well as
the recommended final vehicle classification scheme. The
purpose of both classification schemes is to divide each
manufacturer's product line into groups of vehicles which are
expected to emit approximately the same amount of HC vapors
(family) and have the same control system (system). Jince
these six parameters are recommended to be evaporative control
system determinants, modification by secondary manufacturers
would not be permitted without recertification.
This is current Agency policy in the liyht-duty truck
(LDT) area. Some LDTs are sold as incomplete vehicles or are
modified by secondary manufacturers. secondary manufacturers
who modify vehicles so as to remove them from inclusion in the
engine-system, evaporative emission family and/or evaporative
emission control system combination in which the original
vehicle was certified may be subject to the proscriptions
against tampering or selling uncertified vehicles. We
recommend that this policy be retained in this HDG regulation.
Thus, the modification of the six parameters discussed above
(as well as the other evaporative control system determinants
listed in the issue "Certification Procedure") could constitute
tampering and would, therefore, be illegal unless the vehicle
was recertified. However, secondary manufacturers should not
have to recertify because the parameters selected as
evaporative family-system determinants should not have to be
modified. (The upgrading of hydrocarbon storage devices is an
exception as will be discussed below).
The final parameter listed in the nprm for consideration
under the proposed incomplete vehicle requirements is "fuel
tank volume." Unlike the other parameters which we have
recommended to delete in this final rule because their impacts
on evaporative emissions are considered to be minor, fuel tank
volume is known to have a major impact on evaporative
emissions. in its comments to this rulemaking, CM claimed that
every 10 gallon increase in fuel tank volume results in a u.10
gpt increase in controlled evaporative emissions. in the uprm

-------
-54-
we presented data showing only a 0.05 gpt increase in
evaporative emissions with a 10 gallon increase in fuel tank
volume. Although there is a substantial difference between
Givi's and our claims, both are significant when it is noted that
HDGs can have fuel tank volumes of up to 150 gallons. If a
secondary manufacturer added fuel tank capacity to bring the
total from 40 gallons to 150 gallons, GM's data would predict a
1.98 gpt increase in controlled emissions while our data would
predict a 0.55 gpt increase. Both increases could cause a
vehicle to emit above the level of the standard. Another
factor to remember is that these increases assume that the
working capacity of the HC vapor storage system and the purging
system are adequate to handle the increase in uncontrolled HC
emissions resulting from the fuel tank volume increase. if the
working capacity was not sufficient, breakthrough would occur
causing much higher in-use emissions than those predicted
above. If the purging system could not adequately purge the
storage devices, then those devices would eventually become
overloaded causing breakthrough also. It is clear that the
problem of increased fuel tank volume should not be dismissed
as minor.
Discussions with the primary manufacturers have revealed
that a large majority of orders for incomplete vehicles include
permanent fuel tanks of a specified volume. Chrysler and Id
informed us that all of their incomplete HDGs have permanent
fuel tanks installed when they leave the factory. Ford said
that only about 10 percent of its over 16,000 lbs. GVWR
vehicles (all of which are sold incomplete) do not have
permanent fuel tanks. Instead, these vehicles have a small,
temporary fuel container used to raove the vehicle through the
delivery system. GM told us that only about 10 percent of its
HDGs less than 14,000 lbs. GVWR are sold with incomplete fuel
systems.
Thus, there are relatively few incomplete HDGs that leave
the factory with the final fuel tank volume of the completed
vehicle unknown. In addition to the few HDGs that leave the
factory without permanent fuel tanks there are also some HDGs
that will have an extra fuel tank(s) added to the permanent,
factory-installed fuel tank{s). This might occur, for example,
in the case of a dealer who buys a chassis/engine combination
without having a customer for it. When the customer does
materialize he might want an additional amount of fuel
capacity. vie do not know the total number of instances where
HDGs have fuel tanks added after they have left the factory bat
we believe the number to be small. We believe the primary
manufacturer can, with relative ease, make a good faith e^iioct
to seek and find most of these situations and inform trie
potential dealers/customer that the evaporative emission
control system must be adequate to handle the HC emissions lcou
the vehicle in its completed configuration. Once tne
dealers/secondary manufacturers become aware of the

-------
-55-
requirement, we expect that they will inform the primary
manufacturers of the total fuel tank volume that they foresee
the completed vehicle having. The primary manufacturer will
then supply the appropriate evaporative emission control system
with the incomplete vehicle. Therefore/ we recommend that fuel
tank volume be retained as a parameter which must be considered
for certification of incomplete vehicles.
It should be noted that "fuel tank volume" is not an
evaporative family-system determinant. It was a family
determinant in the proposed vehicle classification system,
however, we are recommending that it be deleted in the final
rule in order that the primary manufacturers can have greater
flexibility in grouping their product lines. For example, one
manufacturer might choose to have a wide range of fuel tank
volumes controlled by a single evaporative emission control
system which would imply overdesiyn for the lesser fuel tank
volume vehicles whereas another manufacturer might choose to
design separate control systems for each fuel tank volume he
sells thereby having no overdesign but having, instead, more
family-system combinations to certify.
To summarize, the primary manufacturer will place each
incomplete vehicle he wishes to sell into an evaporative
emission family-evaporative emission control system
combination. The family determinants will be the method of
fuel/air metering (i.e., fuel injection vs. carburetion) and
the carburetor fuel bowl volume (within a 10 cc range). The
control system determinants will be the method of vapor
storage, the method of carburetor sealing, the method of air
cleaner sealing, the vapor storage working capacity (within 20
grams), the number of storage devices, the method of purging
stored vapors, the specifications of the purge system, the
method of venting the carburetor during both engine off and
engine operation, the liquid fuel hose material, and the
configuration of the storage system for fuel tank emissions.
Manufacturers must certify that each family-system meets the
standards as built, up to a stated maximum fuel tank capacity.
For incomplete vehicles, the primary manufacturers can use a
typical frontal area in the road load equation for
determination of the dynamometer horsepower setting. The large
majority of incomplete vehicles will have permanent fuel tanks
installed at the factory. These incomplete vehicles can be
certified with that fuel tank volume. For incomplete vehicles,
the primary manufacturer will include a label which states that
the evaporative control system that is supplied with the
vehicle was designed to handle a specified maximum fuel tank
volume.
Situations where the secondary manufacturer will add fuel
tank volume beyond the maximum specified by the primary
manufacturers are expected to be few. As discussed above, that
portion of all incomplete vehicles which leave the factory

-------
-56-
without permanent fuel tanks is small and, for the majority of
them, the primary manufacturer will supply a control system
designed to handle the emissions of the fuel tank(s) when the
HDG is completed, secondary manufacturers who order incomplete
HDGs and install fuel tanks should tell the primary
manufacturers what amount of fuel tank capacity they plan to
install. The primary manufacturers will then supply a control
system large enough to handle that fuel tank capacity.
However, there will still be some instances where a
secondary manufacturer may want to increase the fuel tank
capacity beyond the maximum specified by the primary
manufacturer. For example, the secondary manufacturer might
not know the ultimate desired fuel tank capacity until after
the incomplete vehicle has been purchased. This situation
might occur where a secondary manufac turer buys an
engine/chassis combination without knowing who the ultimate
owner might be. Information available to the Agency does not
indicate that these "custom-made, third party" HDGs are common
and the few that do occur will niost probably require total fuel
tank capacities which are within the limit set by the primary
manufacturer.
For those few instances where a secondary manufacturer
wishes to exceed the primary manufacturer's fuel tank volume,
the secondary manufacturer will be required to increase the
adsorption capacity of the hydrocarbon storage device(s)
according to the following formula:
Capf = Final amount of fuel tank vapor storage material
in the hydrocarbon storage device, grams.
Capj_ = Initial amount of fuel tank vapor storage
material in the hydrocarbon storage device as .supplied by the
primary manufacturer, grams.
T. Vol. = Total fuel tank volume of vehicle when
completed, gallons.
Hax. Vol. = Maximum fuel tank volume as specified by the
primary manufactucer, gallons.
The above equation increases the capacity of the
hydrocarbon storage device in proportion to the increase in- tne
fuel tank volume. This is being done because the increase in
emissions is essentially proportional to the increase in fuel
tank volume. Therefore, we have concluded that the resultant
increase in vay^r storage material will be sufficient to
control the increased vapors from the larger fuel tank.
Where:

-------
-57-
If the secondary manufacturer needs to add a second
canister rather than just change to a larger canister than that
supplied by the yriitiary manufacturer, the secondary
manufacturer shall hook up the new canister in series after the
first one and the first canister must be sealed to eliminate
any openings to the atmosphere. Also, the second canister's
elevation shall be equal or higher than the first canister's.
Any fuel vapor hosing used by the secondary mamifacturer must
be at least as impermeable to hydrocarbon vapors as that used
by the primary manufacturer and the vapor storage material Kiust
have the saae adsorptive characteristics as that used by the
primary manufacturer,
A secondary manufacturer who wishes to add fuel tank
capacity beyond the maximum specified by the primary
manufacturer will be required to submit a written statement to
the Administrator that it has upgraded the hydrocarbon storaye
device(s) according to the requirements discussed above.
1'his discussion will now turn to the question of
liability. The primary manufacturers were concerned that this
regulation, as proposed, would leave them liable cut
modifications and/or additions to incomplete vehicles which
caused those vehicles to be in noncompliance with the
standard. This Final Rule will clarify this point. Each
incomplete vehicle will be certified as meeting the stanuards
as built, at sowe GVWR, at some frontal area and with sone
maximum fuel tank volume. Also, these vehicles will all have
evaporative emission control systeras designed to contain the HC
vapors emitted. If a secondary manufacturer adds fuel tank
volume above the primary manufacturer's specified maximum, or
changes any other parameters, then, potential liability for
noncompliance will shift to the secondary manufacturer if tne
change is the cause of the noncompliance.
The principal arena for questions of liability is the area
of in-use emissions. Although, EPA has the option of doin^j
i;\-use testing of HDGs for evaporative emissions (see the issue
"Certification procedure"'}, we do not have plans for such
testing. However, i£ EPA does in-use testing, it would
normally be limited to situations where a number of HDGs are
suspected of evaporative emissions well above the standard.
Since the parameters that secondary manufacturers would be
to modify (without recertificacion} ace expected to hove a
minimal af£ect on evaporative emissions (e.y., body cy.je,
relocation of exhaust, frontal area), the need fcr u
comprehensive in-use testiny program would most probably be t-.ie
fault of a prinary manufacturer's inadequate design ur
component defects and excess emissions would show up in a ¦.vivii
variety of secondary configurations. Should this not pro--
be the case, then the Agency would fully investigate; c-o
determine appropriate placement of liability. If tne pc li:™:/
manufacturers snake a good faith effort in uesiyniu^

-------
-58-
producing HDG evaporative emission control, systems and if they
properly transfer the limitations and assemblage instructions
of those systems to the secondary manufacturers/ then EPA
should have little reason to exercise its option to do in-use
evaporative emission testing of HDGs.
Recommendat ion
We recommend that the primary manufacturer place each of
its incomplete vehicles in an evaporative emission
family-control system grouping as defined in the issue
"Certification Procedure." Each incomplete vehicle will have a
label on it specifying the maximum fuel tank volume that the
control system was designed for. Secondary manufacturers will
be responsiole for correct assembly of the evaporative emission
control system (if applicable). Secondary manufacturers will
be subject to tampering regulations if they modify the control
system so as to remove the vehicle from its original, certified
family-system combination. If a secondary manufacturer wishes
to add fuel tank volume in excess of that specified by the
primary manufacturer, then it must submit a written statement
to EPA that it has upgraded the hydrocarbon storage device(s)
as required.

-------
-59-
References
1. Chrysler Corp., etal vs. EPA, 600 Fed. 2d. 904
(1979) .

-------
-60-
C.. Issue: Technical Feasibility
Summary of the Issue
In the NPRM EPA proposed an evaporative emission standard
for gasoline-fueled heavy-duty vehicles (HDGs) of 3.0 grams per
test (g/test). The preamble to the proposed regulations
included EPA's rationale in support of the technical
feasibility of the proposed standard. That discussion
examined: 1) the level of evaporative emission control for
light-duty trucks (LDTs),. and 2) the differences between LDTs
and HDGs with regards to evaporative emissions. One major
difference is that maximum fuel tank capacities are greater for
HDGs. While LDTs usually don't have maximum fuel tank
capacities exceeding 40 gallons, HDG fuel tank capacities can
exceed 100 gallons. A regression analysis of 1979 model year
LDT certification data predicted (although these was not a
strong correlation) that a 10 gallon increase in fuel tank
volume results in a .05 g/test increase in total controlled
evaporative emissions. Thus, a 100-gallon fuel tank would have
0.3 g/test higher evaporative emissions than a 40 gallon fuel
tank.
Additionally, the preamble cited data gathered by the
American Petroleum Institiute (API) as part of a refueling loss
study which demonstrated an efficiency of 99.5 percent in
controlling evaporative losses. Since both industry and E?A
have shown that an uncontrolled, 100-gallon fuel systein
generates about 50 grams of HC during the diurnal part of the
evaorative emissions test, the API result indicates that a
100-gallon fuel tank can be controlled to a level of 0.23 grams
(50 grams HC x (100% - 99.55%)).
Another difference between LDTs and HDGs is that nonfuel
or "background" evaporative emission levels are expected to be
higher for HDGs. Sources of background HC emissions include
paint, tires, sealers and sound deadeners. since HDGs are, on
the average, larger than LDTs, the amounts of such background
sources of HC are also larger. For example, payload boxes on
trucks such as large delivery vans use. more paint than the
payload area of a pickup truck. EPA tested a new 1977 Ford
truck for background emission levels and found that it emitted
about 0.6 g/test. Ford tested a LUT and found that it emitted
0.31 g/test. These results indicate that HDGs emit about 0.3
g/test more background emissions than do LDTs and the proposed
standard for HDGs accounted for this.
A third difference between LDTs and HDGs that affects
evaporative emissions is the volume of the carburetor fuel
bowl. The main source of HC emissions during, the hot-soak
portion of the SHED test is the carburetor fuel bowl.
Increasing fuel bowl volume and increasing peak temperatures
both contribute to increasing hot-soak losses. With regaru to

-------
-61-
temperature, data submitted to EPA by Ford indicated that
higher engine output, as required in the HDG driving schedule
as compared- to the LDT driving schedule, does not result in
higher carburetor gasoline temperatures during the hot-soak
test. So LDTs and HDGs see about the same carburetor fuel bowl
temperatures. Typically, LDTs have fuel bowl volumes ranging
from 70 cc to 150 cc. We believed that HDGs had maximum fuel
bowl volumes of 268 cc. One 1979 LDT family also had a
carburetor fuel bowl volume of 268 cc. Three of these LDTs
with a carburetor fuel bowl volume of 268 cc were tested for
evaporative emissions and the best of these three tests was 2.2
g/test. This data demonstrated the feasibility of controlling
evaporative emissions from vehicles with carburetor fuel bowl
volumes representative of large HDGs to a level below 3.0
g/test.
Using the above worst case situations, EPA determined that
a 3.0 g/test standard was technically feasible. A LDT with a
large carburetor fuel bowl volume (268 cc) emitted 2.2 g/test.
If this vehicle had a total fuel tank capacity of 100 gallons,
then it might have emitted an additional 0.3 g/test.
Furthermore, if this LDT was physically as large as the bigger
HDGs, then it might have emitted an additional 0.3 g/test in
background emissions. The summation of the above worst case
conditions yields a total of 2.8 g/test. Therefore, the
proposed HDG evaporative emission standard not only appeared to
be technically feasible but allowed for a safety margin as well.
Summary of Comments
The four primary HDG manufacturers (GM, Ford, Chrysler,
and IH) were the only commenters on this issue. In general,
the commenters felt that a 3.0 g/test standard could be met for
the lower weight classes of HDGs although they differed as to
what weight range to use. GM stated that its HDGs of 12,000
lbs. GVW or less could meet the proposed standard. Ford
commented that the 3.0 g/test level would be appropriate for
its 14,000 lbs. GVW and under HDGs. IH stated that the
cut-point should be 16,000 lbs. GVW. IH does not produce any
HDGs in the lower weight classes. Chrysler did not discuss what
level of control its HDGs could meet. Instead, Chrysler
claimed that EPA's rationale and data used to show the 3.U
g/test standard to be technically feasible was inadequate.
GM and Chrysler specifically attacked EPA's technical
feasibility rationale as given in the preamble of the i:p ;<[•!.
Doth manufacturers claimed EPA's conclusion that HDGs would
emit about 0.3 g/test more than LDTs because of increased ::uei
tank capacities was invalid. GM SHED tested two HDGs: one
with a fuel tank capacity of 40 gallons and one with a ii:el
tank capacity of 100 gallons. The difference between t:ie
diurnal portions u £ the two tests was 1.06 grams of HC. G:i
concluded, therefore, that a 10-gallon increase in fuel tank

-------
-62-
capacity would result in a 0.18 gram increase in diurnal test
emissions rather than the 0.05 gram increase that EPA had
concluded. Chrysler claimed that a report[l] by EPA showed
that losses are not directly proportional to the volume of the
fuel tank and, therefore, there must be other tank factors
besides volume contributing to the diurnal losses. From this
Chrysler concluded that EPA's estimate that a 100-gallon fuel
tank would emit 0.3 g/test more than a 40-gallon tank was not
valid. Furthermore, Chrysler stated that the API refueling
study is irrelevant and may have been biased because the API
wanted to show that on-board refueling loss control systems
would be better than stationary recovery systems.
GM and Chrysler both claimed that EPA's estimate of the
difference in background emissions between HDGs and LDTs is
invalid. GM tested background levels of two HDGs and found
significantly higher background emission rates than EPA's
estimate. GM's background data on these two trucks (both 1-1/2
years old and thoroughly cleaned) are as follows:
Diurnal	Hot-Soak	Total
Background	Background	Background
Vehicle	(g)	(g)	(g)
Chevrolet Pickup .18	.60	.78
GMC Tractor	.86	1.46	2.32
GM went on to estimate that 95 percent of the pickup's and 55
percent of the tractor's background emissions were due to vapor
permeation through liquid fuel hoses. GM recommended that an
EPA/Industry work group be formed to find a better solution to
the background emissions problem.
GM and Chrysler claimed that EPA's estimate of background
emissions was based on insufficient data because only one
heavy-duty truck was tested. Furthermore, Chrysler stated that
there was no indication that the one truck tested was a worst
case situation and GM stated that since the truck was a 1977
model, it was not state-of-the-art. Chrysler recommended that
background levels be measured and then subtracted from
certification test results.
GM, Chrysler and IH questioned the technical feasibility
of controlling hot-soak losses from carburetors with large fuel
bowl volumes. GM stated that while EPA had claimed that the
maximum fuel bowl volume for HDGs was 268 cc, it markets a
carburetor with a fuel bowl volume of 400 cc. Thus, EPA has
not shown feasibility for this worst case. Chrysler stated
that EPA's use of the lowest test result from the three tests
of 1979 LDTs with fuel bowls of 268 cc was unfair. The lowest
result was 2.2 g/test but the other two tests yieldeu 2.7
g/test and 4.5 g/test. IH concluded from its HDG evaporative
testing program (13 tests) that "four-barrel engines may be

-------
-63-
rnore difficult to control than two-barrel engines. This may
suggest that the 3 gram standard may be too stringent for a
four-barrel engine." IH did not indicate why four-barrel
engines may be harder to control but larger fuel bowl volumes
may be a contributing factor.
In the preamble to the .NPRM, EPA presented certification
data which shewed that LDTs could be controlled to less than
3.0 g/test. GM stated that while 69 of the 137 LDT evaporative
tests for 1979 model year certification had results as low as
2.2 g/test, the remaining 68 vehicles averaged emissions of
3.56 g/test. Thus, GM claimed that rather than proving
evaporative emissions could be controlled below 3.0 g/test for
LDTs, this data showed that a large percentage of LDTs, using
the best available technology in control hardware, failed to
meet a 3.0 g/test standard.
GM and Chrysler both claimed that EPA's use of light-duty
data and experience to predict HDG evaporative emissions is
invalid. Chrysler stated that since the proposed HDG driving
cycle and test procedure are Different than the light-cuty
driving cycle ar.o test procedure, LDT test results are not
equivalent to future HDG test results and thus cannot be used
to predict HDG evaporative emissions. GM cited EPA as stating
in 44 Federal Register at page 46298 , column 2 (August 7, 1S79)
that: "Large heavy-duty trucks have characteristics different
from those of trucks uncer 8,500 lbs." This statement referred
to evaporative emissions. GM then stated that in spite of the
above statement, EPA usee LDT test methods and projections of
LDT test results to preoict HDG evaporative emissions. In GM's
opinion there is no technical rationale "which supports EPA's
new and inconsistent treatment of (HDGs)."
Since EPA allegedly did not adequately demonstrate the
technical feasibility of the proposed standard, Chrysler stated
the Agency has not fulfilled its statutory obligation to do so,
found in the Clean Air Act, §202 (a)(2) and §202 (b) (1) (C) . GM
claimed that EPA failed to live up to its obligation to
consider the representativeness of the test data used to
develop and justify the standard. GK claimed that this
obligation was recently reaffirmed in National Lime Association
vs. EPA, D.C. Circuit, No. 78-1385, May 19, 1980.
GM also suggested that backg round emissions will cause
production HDGs to exceed the stancara at the time of their
introouction into commerce and that the legal implications of
this should be aocresseo by EPA. GM stated that although the
preamble indicatec that a n.anuf actur er can reduce backgrcunc
emissions for HDG certification testing in similar ways as they
are allowed tor in light-duty evaporative emission testing, the
regulations do not r.er.tion this.

-------
-64-
Ford did not discuss EPA's technical feasibility
rationale. Instead Ford conducted an extensive HDG testing
program. Ford tested four HDGs a total of 126 times. This is
far more testing than done by either GM (5 tests) or IH (13
tests). Table 1 is Ford's testing summary sheet. Ford also
ran a total of three background tests on a HDG. From its
testing results Ford concluded the following three points:
1.	The current capability of the existing California
control system, using the proposed test procedure, is about
3.0-4.0 g/test based on testing of hiyh voluiae trucks which are
relatively lightweight.
2.	Because the emission level is a function of the test
procedure used, a new development program would be required to
modify ' the existing Ford/California systems to assure
compliance with the proposed Federal 3.0 g/test standard.
Anticipated areas of need improvements are:
a.	Design of new purge control systems
b.	Carburetor bowl vent design changes
c.	Fuel system vapor integrity improvements
3.	Background HC levels on aged HDG vehicles are most
probably in the area of 0.5 g/test.
Ford's technical comments describing the effect of
different vehicle inertia weights and driving cycles will be
discussed in the next section, "Analysis of Comments". In
general, Ford did not indicate the proposed 3.0 g/test standard
would be infeasible but rather that the control of HDGs to this
level would not be possible by 1983 due to leadtime and
facility problems.
Analysis of the Comments
In general, the commenters agreed with EPA that a 3.0
g/test standard will be feasible for the lighter HDGs. GM
stated "We recommend that a 3 g/test standard be implemented
for (HDGs) below 12,000 GVW. We believe that the tecnnology
does exist to meet the three grain limit on the smaller HDVs
which have background rates more closely related to LDTn . Ford
stated that based on its test program "The current capability
of the existing California control system, using the proposed
test procedure, is about 3.0-4.0 y/test based on testing of
high volume trucks which are relatively light-weight" . The
HDGs Ford tested had inertia test weight (IW) settings •..¦jiich
ranged from 6250 lbs to 3G,00Q lbs. Since the final test
procedure calls for IW settings equal to 50 percent of gvw, we
presume that the HDGs Ford tested were representative oi ::jgs
with GVWs of 12,500 lbs to 72,000 lbs. Ford also stated that
it "believes that the present evaporative control systems for
LDT could be extended to 8500-14 ,000 lb GVW vehicles by 19 d 3

-------
Table 1
Ford's Vehicle Test Results summary
Vehicle Number
Test Numbers

Route

IW & HP
Setting
Average
Grams/Test
C-700
#402
8 thru 14 (7)


Heavy-Duty

19,250/59.6
2.99
Total
Tests
16 thru 18, 2 0,
&
21 (5)
Heavy-Duty

19,250/88.7
3.31
Run =
36
22 thru 25 (4)


Heavy-Duty

36,000/110
4 .54


28


Heavy-Duty
Dual
36,000/110
5 .30


32, 33, 36 (3)


Heavy-Duty
Dual
19,250/88.7
3.02
C-700
#411
9 thru 14 (6)


Heavy-Duty

19 ,250/59 .6
2.71
Total
Tests
17, 18, 20, 21,
22
(5)
Heavy-Duty

19,250/88.7
3.15
Run =
39
24 thru 28 (5)


Ileavy-Duty

36,000/110
4.80


29, 30, (2)


Heavy-Duty
Dual
36,000/110
3 .23


35, 36, 39 (3)


Heavy-Duty
Dual
19,250/88:7
4.02
F-350
#414
11, 12, 14 (3)


Heavy-Duty

7 ,000/53 .3
2.98
Total
Tests
15 & 16 (2)


Heavy-Duty

7,000/24.0
2 .65
Run =
28
2b thru 28 (3)


Light-Duty

7,000/23.9
2.12
F-350
#415
6 thru 8 (3)


Heavy-Duty

6 ,250/22 .9
2 .74
Total
Tests
10, 12 thru 14
(4)

Light-Duty

6,250/23 .4
2.21
Run =
23
15 thru 19, 22,
23
(7)
Heavy-Duty

7 ,000/57 .5
3.51

-------
-66-
and Ford probably could recalibrate the system to meet the
proposed 3 g/test standard with the associated test procedures."
IH tested three HDGs and concluded that "four-barrel
engines may be more difficult to control than two-barrel
engines. This may suggest that the 3 gram standard may be too
stringent for a four-barrel engine." This is the only comment
IH gave on what it considers to be a technically feasible level
of control. Since IH does not produce HDGs in the lower weight
classes (below Class VI), this discussion of technical
feasibility for the lower weight HDGs does not apply to it.
However, we will refer to IH's comments later in the discussion
of the higher weight HDGs.
Chrysler's comments attacked EPA's derivation of the
technical feasibility of the proposed 3.0 g/test standard but
Chrysler did not indicate what level it believed was feasible.
Since all of Chrysler's HDG production is in the lower weight
classes (less than 16,000 lbs), the level of the standard for
the lower weight riDGs is important to it.
GM and Ford both agree that the 3.0 g/test level of
control is technically feasible for the lower weight classes
of HDGs. LDTs, which go up to 8,500 lbs., are currently
controlled to a level of 1.3 g/test (1981 certification testing
average). The differences between LDTs and the lower weight
classes of HDGs are relatively minor since typically these HDGs
have similar fuel tank capacities (although somewhat laryer),
similar background emissions and, in most cases, they use
carburetors which are also used on LDTs. A 3.0 g/test standard
for HDGs would be 50 percent higher than the current LDT
standard of 2.0 g/test which will more than compensate for
these differences. Chrysler produces many LDTs and therefore,
possesses the technical experience in controlling LDT
evaporative emissions to the 2.0 g/test level. In light of the
above, Chrysler should have no trouble controlling its HDGs,
which are all relatively light weight, to the same level that
GM and Ford have agreed to, that is, 3.0 g/test.
The above discussion verified that the proposed standard
of 3.0 g/test is technically feasible for the lighter weight
HDGs. However, the commenters disagreed somewhat as to the
exact weight level at which the 3.0 g/test standard would no
longer be appropriate. GM commented that the split between
"light" HDGs and "neavy" HDGs should be 12,000 lbs GVW' while
Ford claimed it should be at 14,000 lbs GVW. Chrysler did not
comment on this but IH mentioned 16 ,000 lbs GVW as a yood
breakpoint.
From Ford's and GM's comments it is clear that the
breakpoint should be in the range of 12,000 lbs GVW to 14,uuU
lbs GVW. GM gave no reason why it had chosen 12 ,000 lbs GVW.
Ford did not give a reason either but its choice of 14 ,000 Ids

-------
-67-
GVW coincides with the breakpoint between heavy-duty Classes
III and IV and we assume that is why Ford chose 14,000 lbs
GVW. Maximum fuel tank capacities and vehicle sizes (as they
relate to background emissions) both gradually yet larger with
increasing GVW but there is much overlap between HDGs in the
narrow weight range of 12,000 to 14,000 lbs. In fact, it is
highly unlikely that any discernible difference in these
parameters could be detected even if an exhaustive study were
undertaken comparing HDGs of 12,000 GVW to those of 14,000 lb
GVW. Therefore, we will make the split at 14,000 lbs GVW so as
to coincide with the widely accepted HDV class groupings.
Thus, all HDGs in Class III and below (less than or equal to
14,000 lb GVW) will be required to meet the standard of 3.0
g/test.
We now turn our attention to those HDGs with GVWs greater
than 14,000 lbs GVW but less than or equal to 26,000 lbs GVW
(Classes IV-VI) (HDGs with GVWs greater than 26,000 lbs will be
certified by engineering evaluation.) These heavier HDGs will
be more difficult to control because the three major factors
contributing to evaporative emissions reach their respective
maximums with increasing GVW. Maximum fuel tank capacity is
expected to be around 150 gallons. This amount of fuel is
needed on large HDGs where fuel consumption is highest to yive
such vehicles sufficient range. HDGs carrying the bigyest
loads require the biggest engines which in turn require
carburetors with the maximum fuel bowl volumes. Finally, in
many cases, those HDGs with the most painted surface area (and,
therefore, the highest background emissions) are the largest
HDGs both in terms of physical dimensions and GVW.
Our analysis of the feasibility of a 3.0 g/test standard
for these "heavy" HDGs is positive but only marginally so.
That analysis is based on extrapolation of light-duty increases
in controlled emissions versus increases in fuel tank volume
and carburetor fuel bowl volume. The only data on actual HDG
evaporative emissions testing we have is that which the
commenters supplied.
GH tested one relatively large HDG. The result of this
test was 6.38 g/test. This vehicle was equipped with the 1979
California evaporative control system and no attempt was wade
to improve on that design. Two obvious improvements would ue
to use a less permeable liquid fuel line and to xiicluue
activated charcoal in the air cleaner. Both of these control
system strategies are standard light-duty technology.
According to GM, the liquid fuel lines emitted 1.27 g/test ^:iu
the addition of a carbon air cleaner might reduce the tonal
test result by another 0.3 g/test. We have no doubt tint
further improvements could reduce the evaporative emissions
from this HDG to below 4.0 g/test. GM states on page 15 ci; izs
final, written comments, "We recommend that a 3g/test standard
be implemented for GFHDV below 12,000 GVW and a standard of 4

-------
-68-
g/test for GFHDV above 12,000 GVW. * Later, on page 16, Gto
states "In all of our previous responses to the EPA, we have
always recommended a standard of 4 g/test for GFHDV."
therefore, a level of 4.0 g/test is appropriate for GM.
Ford did the most testing by far. It tested HDGs with
inertia weight settings of 6,250 lbs to 36,000 lbs. At the
test weight requirement of 50 percent of GVW, these vehicles
represent a GVW range of 12,500 lbs to 72,000 lbs which
encompasses the range of GVWs to which the 4.0 g/test standard
will apply (14,001 lbs to 26,000 lbs). Ford tested two, C-700
HDGs at an inertia weight setting of 19,250 lbs (i.e., a GVW of
38,500 lbs). The average result of these 23 tests was 3.02
g/test. This clearly shows the feasibility of the 4.0 g/test
standard especially when it is remembered that these HDGs had
little, if any, R&D. Ford states in its final comments that
"the proposed 3 g/test standard for 1983 cannot be attained
using the proposed test procedure because of lead time/facility
issues." Thus, Ford indirectly admitted that the 3.0 y/test
standard is technically feasible not only for the "light" HDGs
but for the "heavy" HDGs as well. Ford also stated that based
on its testing program, "the current capability of the existing
California control system, using the proposed test procedure,
is about 3.0-4.0 g/test ...." It is clear, therefore, that not
only is a 4.0 g/test standard appropriate for Ford, it is one
that Ford's California control system is meeting already.
IH tested three HDGs, one 2-bbl and two 4-bbl. The 2-bbl
was tested as a 25,440 lbs GVW vehicle and clearly would meet a
4.0 g/test standard since the test results were 1.69 g/test and
1.67 g/test after the installation of new tank gaskets and
screws. Vehicle #341, a 4-bbl had a lowest test of 4.51 g/test
and vehicle #342, another 4-bbl, had a lowest test of 3.30
g/test. IH's only comment on what it thought was a feasible
level of control was "We would conclude that four-barrel
engines may be more difficult to control than two-barrel
engines. This may suggest that the 3 gram standard may be too
stringent for a four-barrel engine." We agree with IH and
since all of IH's HDGs would be subject to the 4.0 g/test
standard, we conclude that the 4.0 g/test level is appropriate
for IH also.
Chrysler does not produce any "heavy" HDGs but if it were
to begin such production we would presume that it, too, could
meet the 4.0 g/test standard since GM, Ford and IH can.
Recommendation
We recommend that the proposed standard of 3.0 g/test be
retained for HDGs with GVWs between 8 ,500 lbs and 14,001 Ijs
(Classes IIB and III). For HDGs with GVWs greater than 14,Uu«j
lbs, we recommend a standard of 4.0 g/test.

-------
-69-
References
1. "Heavy-Duty Truck Evaporative Eiaissions Regulation
Development," EPA Technical support Report for Reyulatory Action,
John Corcoran, July 1976.

-------
-70-
D. Issue: Leadtime
Summary of the Issue
In the NPRM, we determined that 1983 was the earliest
feasible model year for implementation of this regulation.
This determination was in accordance with sections 202(a) and
202 (b)(1)(C) of the Clean Air Act. The time required to
implement this regulation was divided into three categories:.
1) time for research and development (R&D) to identify
effective control components and systems, 2) time to finalize
production designs and produce the necessary drawings, and 3)
time to effect tooling changes for production of the new
components. We estimated that tooling changes would take 10
months and that the time to finalize production designs and
produce the drawings would be 6 months. Thus, a total of 16
months would be needed from the time the control components and
systems had been identified to the time of their introduction
on the new models. Since we assumed that the HDG model year
begins September 1 of each year, the identification of the
control components and systems would have to be completed by
Hay 1 of the preceding model year.
The NPRM was published on April 30, 1930 and the Final
Rule was projected to be published in December of 1980. The
time between Final Rule publication and May 1981 was,
therefore, projected to be about 6 months. We estimated that
this would be enough time to identify the necessary control
components and systems because those components and systems
were expected to be virtually the same as the current LDT
components and systems. In fact, the majority of HDG
carburetors are also used on LDTs and most of the HDGs in the
lower weight ranges have fuel tank capacities similar to LDTs.
Thus, the R&D for HDGs was expected to be quite limited because
the evaporative control systems for most HDGs would be the same
or slightly modified versions (e.g., bigger canisters, higher
purge rate) of current LDT systems.
The largest HDGs were expected to require most of the R&D
work. Vehicles with carburetors not in use on LDTs and with
fuel tank capacities in the 100 gallon plus range were expected
to need R&D test work. We recognized that some manufacturers
might not have a teat cell operational by the projected date of
Final Rule publication. However, manufacturers had indicated
that component tests gave results very close to full-SHED test
results. Thus, manufacturers could use alternative test
procedures to beyin the limited R&D work needed for the larye
HDGs until their full-SHED facilities were completed. We
concluded that all manufacturers would be able to identify the
new control components and systems by May 1, 1981 and,
therefore, the earliest feasible model year was projected to oe
1983 .

-------
-71-
Summary o£ the Comments
The four primary manufacturers of IIDGs (GM, Ford,
Chrysler, IH) were the only commenters on this issue. In
general these manufacturers disagreed with us that there is
adequate leadtime for 1983 model year implementation. Since
each manufacturer's projected timetable is different, we will
present their comments and projections separately.
GM presented a detailed leadtime chart which estimates
that 35 months would be required after publication of the final
rule. GM claimed that a new test facility with required test
equipment would be needed. Corporation appropriations would
require 4 months. Then, design and construction of the test
facility along with procurement and installation of testing
equipment was projected for the next 20 months. Finally,
system checkout was scheduled to take 5 months making the total
leadtime required to complete the new test facilities 29 months.
Concurrent with facility construction would be control
component and system development. A total of 29 months is
estimated for this development work with no full SHED testing
until the fourteenth month after final rule publication. At
that time component, background and purge system testing begins
and is projected to require 5 months. Next, certification
fleet assembly and new control system development testing are
scheduled to be completed by the twenty-fifth month at which
time durability testing of the new components and systems is
scheduled to begin. Durability testing is projected to take 4
months.
After 29 months, the certification facilities and the
system development work are projected to be complete. Next, GM
estimated three months for certification testing and two months
for EPA review. GW's timing chart shows a 1-month delay after
EPA's review before the start of production. GM did not
provide an explanation for this delay.
GM did not agree With the NPRM that the only change to
current carburetors (not already in-use on LDTs) woud be
venting to the charcoal canister. GM claimed gasket material
and stem ana/or rod sealing changes would have to be
evaluated. GM also stated that the NPRM's assumption ol test
facility existence soon after final rule publication is
erroneous. GM claimed that EPA has failed to provide an
acceptable option for SHED testing by manufacturers who do nut
have the facilities. Finally, GM stated that a 'carburetor
design that has undergone significant changes, for this
evaporative emission regulation will be required to unuec.jO
more design changes (and corresponding tooling changes) lur
compliance with the 1934 MY exhaust emission standards.

-------
-72-
Ford also presented a leadtime chart, although it did not
include much detail. Ford claimed that 1800 development tests
and 250 certification tests would be required each model year.
From experience, Ford concluded that slightly more than two
full-SHED test cells would be required. Since Ford already has
one test cell, one or two more would need to be purchased and
installed. Ford projected that 17 months would be required for
this test cell procurement and installation. This estimate
presupposed that EPA would approve worst case testing (e.g.,
testing a given evaporative system with dual fuel tank
application would be sufficient to approve a single tank
application with the same evaporative system without testing).
While the additional test cells are being procured, Ford
would begin control system development with the one test cell
it already has. Ford stated that it anticipated that the
design of new purge control systems, carburetor bowl vent
design changes and fuel system vapor integrity improvements
would be needed. Ford projected that this control system
development would take 29 months based on the availability of
one cell for the first 17 months.
Ford estimated that certification testing	would require an
additional 5 months and EPA review would take	1 month. Ford's
estimate of leadtime, therefore, was 34 months	from the time of
the final rule publication.
Ford also commented on an alternative certification plan
whereby HDGs greater than 14,000 lbs GVW would be certified by
the California method of engineering evaluation and HDGs less
than 14,000 lbs would be certified as proposed in the NPKM.
Ford claimed this alternative could be implemented for the 1983
MY (assuming Final Rule publication by February 1, 1981)
because control system development time would be cut from 29
months to 17 months.
Another concern Ford had was that although its new
heavy-duty engines are introduced in January of each model
year, its new heavy-duty vehicles are introduced in the
preceding September. Ford uses the previous model year's
engines in its new vehicles for the four months until the new
engines are introduced. Thus, a new vehicle model introduced
in September could have a different engine in it by January.
Ford's concern is that, strictly speaking, the NPRM would
require certification of such a vehicle twice, once in
September and again in January.
Finally, Ford stated that it is very much opposed to
installing new evaporative control systems whinh may
necessitate engine recalibration, and therefore, certification
in 1983 (Ford expects engine certification for 1932 and 1933 to
be carryover) only to have to certify its engines again in i9u4
due to the new heavy-duty exhaust emission regulations.

-------
-73-
Delaying the implementation of this regulation for 1 year to
coincide with the implementation of the heavy-duty exhaust
emission regulation would allow Ford to carryover its engines
through 1983.
Chrysler claimed that it would not be feasible to comply
with the proposed regulations by 1983 . Chrysler stated that 12.
months would be needed to purchase and install a SHED anu
dynamometer. Assuming publication of the final rule in
January, 1981, development work could not begin until January,
1982. The remaining time before production of 1983 MY vehicles
was claimed to be insufficient to complete the development of
evaporative control systems .especially in light of the fact
that EPA underestimated the technical effort required.
Additionally, Chrysler stated that evaporative standards
for 1983 MY may impact unfavorably upon the gaseous exhaust
standards. The 1983 evaporative hardware may cause a
recalibration of the 1983 exhaust emission hardware, and then a
recalibration of the evaporative emissions hardware may be
required with the change in 1984 exhaust emission standards.
Evaporative standards for 1983 MY would frustrate any carryover
certification from the 1982 MY for engines while the 1984
exhaust emission standards would frustrate evaporative emission
carryover from the 1983 MY. Chrysler stated that "EPA shoulu
delay the implementation date for at least one year, to the
1984 model year at the earliest. This would allow the industry
and EPA to make a better determination of what is indeed
feasible, and will largely alleviate current leadtime problems."
IH submitted a leadtime chart with its comments. On the
chart, IH projects that procurement of a heavy-duty dynamometer
would take 10 months, installation and checkout of the
dynamometer woud require another 2 months, necessary R&D would
then require 6 months and the resultant carburetor retooling
would need 14 months bringing total leadtime to 32 months. If
the final rule had been published by January 1, 1981 as IH
assumed, then the earliest feasible model year for
implementation would have been 1984.
IH, like Ford, was concerned with the problem of its new
heavy-duty vehicle model year beginning in September while its
new engines are introduced the following January.
The final major area of concern of IH was that currently
all of its heavy-duty enyines are certified by Subpart H. If a
demonstration of compliance for these engines is required in
1983 as a result oi the impact of evaporative emission control
on exhaust emission control, then IH would have to use subpart
D for such engine' certification testing. Ill has planned that
its engines will be certified through 1983 by the carryover
provision. If Hi is required to test its engines using Subpart
D, then they will fail to meet emission standards. ih

-------
-74-
requested that EPA allow use of either Subpart D or Subpart H
to satisfy any engine test requirements that are necessary as a
result of evaporative emission requirements.
Analysis of the Comments
This Final Fule has a number of changes from the NPRM
which affect the leadtime issue. We will first describe these
changes and discuss their influences on leadtiine. Then, we
will discuss each of the four HDG manufacturer's leadtiine
concerns in light of these changes. Finally, we will present
our recommendation of the model year in which to implement this
regulat ion.
The first area of significant change affecting leadtiine is
the procedure by which a manufacturer receives the certificate
of conformity. In the NPRM we proposed that each
manufacturer's product line would be divided into evaporative
emission families- based on three criteria.} 1} method of
fuel/air metering, 2) carburetor fuel bowl volume (within lOcc)
and 3) fuel tank volume (within 20 gallons, or within 25
percent, whichever is greater). This Final Rule deletes fuel
tank volume as a family determinent. This change will
significantly reduce the number of evaporative emission
family-systems because each manufacturer will be free to choose
the range of fuel tank volumes that each control system is
designed to handle. Thus, a manufacturer may find it
advantageous to develop one evaporative control system to cover
a large range of fuel tank volumes even though the control
system may be over-designed for the smaller fuel tank volumes
to which it applies. This change is essentially the worst case
testing that Ford requested in its comments.
The reduction in the number of family-systems means less
research and development work since a manufacturer will only
need to concentrate on the worst case vehicles. Those vehicles
for which assurance of compliance is easy will not need a whole
certification program as they would have under the proposed
certification procedure. Instead, they will be included in an
evaporative emission family-system which also includes the
harder to control worst case vehicles on which the manufacturer
will concentrate his R&D efforts. We estimate that Ford and CM
will be able to reduce the number of evaporative emission
family-systems that they must certify from somewhere in the
high 20's to about 6 or 3. Since the time required for R&D is
a major portion of total leadtime, this change will have a
significant impact on the amount of leadtime needed to
implement the finai rule.
Another reduction in R&D effort will occur due to the
provision that Class VII and VIII EiDGs need only be certiiLeu
by an engineering evaluation. As discussed in the iss^d
"Certification Procedure", we project that virtually all class

-------
-75-
VIII vehicles (GVW greater than 33,000 lbs.) will be powered by
diesel engines by 1984. We also estimate that by 1988 the same
will be true for Class VII vehicles (GVW between 26,000 and
33,000 lbs.). The few gasoline-powered vehicles in these
classes that will be produced do not warrant the cost of a
certification test proyram to ensure that they meet the 4.0 ypt
standard applicable to HDGs between 14,001 lbs. and 26,000 lbs.
GVW. We are confident that extrapolation of the control
technoloyy used on the Class VI HDGs (19,501-26,000 GVW) will
be adequate to control the few Class VII and VIII HDGs to a
level very close to the 4.0 gpt level. Thus, manufacturers
will not have to expend resources and tine on these HDGs except
to install the evaporative emission control systems on them.
The final change to the proposed certification procedure
is that instead of testing all of the evaporative family-system
combinations for compliance, then submitting this certification
test data to EPA and finally, waiting for EPA to review it ana
to issue certificates of conformity; manufacturers will only be
required to submit a statement that each of its evaporative
family-systems meets or is designed to meet the standard. tio
test data or engineering evaluation or explanation will
generally be needed. Elimination of certification testing and
subsequent EPA review will save 5 to 6 months according to the
comments from GM and Ford.
Another important change to the NPRM that will
significantly affect leadtime is the level of the standard.
The manufacturers generally agreed that the existing LDT
control systems could be used for the lighter weiyht HDGs to
meet a 3.0 gpt standard. However, they had concerns that the
heavier HDGs would be more difficult to control to this level
because of increased fuel tank volumes and higher background
emissions. Therefore, most of their R&D effort was expected to
be directed at the larger HDGs meeting a 3.0 gpt standard.
This final rule includes a split standard (as recommended
by GM) of 3.0 gpt for HDGs between 8,501 and 14,000 lbs. GVW
and 4.0 gpt for HDGs between 14,001 and 26,000 lbs. GVW. Thus,
not only have we eliminated the SHED testing for the very lar^e
HDGs (Classes VII and VIII) but we have substantially eased the
standard for the middle classes of HDGs (IV-VI). This change
will decrease the R&D effort even more since the manufacturers
indicated that a 4.0 gpt standard would be easy to meet. (Trie
issue "Technical Feasibility" discusses this in more detail.)
To summarize thus far, this Final Rule includes thr^e
changes to the N?HM that will reduce the amount of R&D time.
First, the number o£ evaporative family-systems will be reduced
substantially allowing the manufacturers to concentrate their
R&D efforts on worst case vehicles. Second, very little if u.iy
R&D effort will be needed for Classes VII and VIII HDGs. Anu
third, the less stringent standard for Class IV-VI HDGs of 4.0

-------
-76-
gpt will be easy to meet with current techno.loyy while the
manufacturers agree that a 3.0 gpt standard can already be met
for the lighter HDGs.
Another change to the NPRM which will influence leadtiue
is that the test procedure no longer requires the manufacturers
to acquire heavy-duty dynamometers. This change is discussed
in the issue "Test Procedure." Instead they will be able to
upgrade existing light-duty dynamometers by the simple addition
of flywheels. By changing the inertia splits to 5U0 lb.
increraents and using a 2,000 lb. trim, an existing light-duty
dynamometer can be converted to provide up to 13,500 lbs.
equivalent vehicle weight (EVW). This, coupled with the change
to the NPRM of test weight being 50 percent of GVW instead of
70 percent, allows existing light-duty dynamometers to be used
for testing all HDGs through Class VI. Therefore, heavy-duty
dynamometers will not be needed since the heaviest HDGs
(Classes VII and VIII) will not need to be tested.
Discussions with test equipment manufacturers, as well as
the comments of the HDG manufacturers, indicate that the
procurement and installation of heavy-duty dynamometers is the
longest of the test facility leadtime requirements (except for
GM which claimed it needed to construct a new building). The
procurement and installation of a heavy-duty dynamometer can
take 12 months as IH commented but the upgrading of a
light-duty dynamometer requires at the most 6 months, which is
also about the time required to install a heavy-duty SHED, the
next longest leadtime item. Thus, the time needed to acquire
testing capability, providing a new building need not be
constructed, will be reduced by 6 months as a result of these
test procedure changes.
Before we present an analysis of each manufacturers'
leadtime position, we will address the concern of Ford and IH
that their new heavy-duty engines are introduced four months
after introduction of their new heavy-duty vehicles. Changes
in a vehicle from one model year to the next can affect
evaporative emissions. For example, the relocation of the
exhaust system could influence the amount of heat that is
transferred to the fuel tank and/or fuel lines. Also, changes
in the shape of the engine compartment or the operators
enclosure might change heat distribution thereby affecting
evaporative emissions. Changes in an engine could also afreet
evaporative emissions. For example, a new carburetor would
need to be tested to assure vapor integrity or a different
engine miyht need a different set of canister jjurgin^
specifications.
Thus, it is possible that a manufacturer miyht have to
certify its new HDGs twice in the same model year. This mi^ht
occur, for example, if a manufacturer introduced a new vehicle
model in September for which a new certificate of conformity

-------
-77-
would be required and then that manufacturer introduced a new
engine model for those new vehicles in the following January.
If the new engine affected the evaporative family-system
combination, then the manufacturer would have to certify the
vehicles again. However, since the manufacturer is always
responsible for the compliance of any new vehicle and/or engine
and the certification procedure for this reyulation is very
simple (ie., a written statement attesting compliance), the
effort involved if a manufacturer did have to certify twice in
one year would generally be limited to
sending EPA a second written statement of compliance. Also,
since heavy-duty vehicles and engines typically do not change
much from year to year, this problem is further reduced. For
the above reasons, therefore, this Final Rule requires that
each new evaporative emission family-system combination be
certified upon its introduction into commerce.
This discussion will now analyze each manufacturer1s
leadtime position. GM claimed that it would need a total of 35
months from the date of final rule publication to the start of
production. The changes in this final rule as they affect r&d
and test equipment procurement would allow GM to substantially
reduce its leadtime requirements were it not for GM's claim
that it needs to construct a new building. In its leadtime
chart, GM allows 4 months to process the appropriations
request, 6 months to design the facility, two months to acquire
quotes and 12 months to construct the facility for a total of
24 months. If GM did need to construct a small building to
house its heavy-duty testing cells, it should not require 24
months to build. Such a building would be a conventional one
v/ith no complex requirements. GM's allowance of 6 months to
design it appears out of proportion. We believe 3 months for
design would be more than adequate. Also, GM allowed 4 months
to process the appropriations request. We feel confident that
with very little extra effort company funds could be
appropriated much more quickly especially since this rulemaking
has been in the works for at least 3 years now. Allowing 3
months for appropriations seems more realistic. Furthermore,
since there is no backlog in the construction industry due to
the severe recession that industry is experiencing (at least in
the Detroit area), time for construction can probably be
reduced by 30 percent (or 3.6 months). To be conservative we
will reduce construction time by only 2 months. Thus, the
construction of the building should take no longer than 18
months.
Simultaneous with the completion of building construct; ion
would be the completion of the installation of' the test
equipment as GM indicated in its comments. GM then allows 5
months for system checkout. Although GM does not explain why L>
months would be needed to check out test equipment that is, ui
most cases, common to light-duty evaporative emissions testing

-------
-78-
and, therefore, very familiar, we presume that because GM
envisioned this facility correlating with EPA's, extra care and
time would be required. Since the final rule does not allow
EPA to confirmatory test or for that matter even require the
manufacturer to do any certification testing, it seer.is
reasonable that system checkout for the purpose of establishing
repeatability would not require more than 3 months.
In its comments GM shows its R&D work being completed at
the same time as the facility system checkout is completed.
Since GM claimed that component test procedures give results
very close to the full.-SHED test, we imply that some of GM's
R&D work will occur coincidently with facility construction.
The f inalikiation of the evaporative emission control systems
can utilize full-SHED testing since such equipment will be
available for use for some months before the facility is
complete. Because the required R&D effort has been
substantially reduced by the changes to the NPKM discussed
above, we conclude that the 21 months allowed to complete the
facility and perform system checkout is also plenty of time to
complete the R&D work. GM then assumed that the next 6 months
would be needed for certification testing and EPA review.
Since the final rule does not include any such requirement,
GM's total leadtime is 21 months even if it constructs a new
building.
Ford stated that 29 months would be required for control
system development based on the availability of 2 test cells
for the first 17 months. The changes in the test procedure as
compared to the NPRM should allow Ford to complete its second
test cell in 6 months instead of 17 since the upgrading of an
existing light-duty dynamometer takes considerably less time
than the procurement and installation of a heavy-duty
dynamometer. Once the second test cell is ready Ford projected
an additional 12 months would be needed to complete the
development work. This seems unlikely since this final rule
has substantially reduced the required R&D. However, we will
assume the full 12 months will be needed. After completing the
control system development work Ford estimated 6 months would
be needed for certification testing and EPA review. Since no
such testing or review is required in this Final Rule, Ford's
total leadtime requirement is 18 months.
Chrysler claimed it would need 12 months to purchase and
install a SHED and dynamometer. This should now only require 6
months as discussed above. The only additional information
Chrysler supplied was that if it had test facilities ready by
January 1982, then it would have enough time to complete the
necessary R&D and certification testing for the 1984 model
year. Thus, we imply that 23 months (i.e., January 1982 -
December 1983) would have been sufficient for Chrysler's k^d
and certification program under the proposed certification
procedure. We can further assume that of this 23 months about

-------
-79-
6 months, were required for certification testing and EPA review
since both GM and Ford indicated as much. Thus, only 17 months
would actually be devoted to R&D work. We will assume that
this can be further reduced due to the decrease in R&D
resulting from changes to the NPRM and to the fact that
Chrysler does not produce HDGs with GVWs nore than about 14,U0U
lbs. This latter fact means that Chrysler will not have to
deal with some of the problems of the large HDGs but instead
can directly apply current LDT technoloyy to all of its HDGs.
Reducing the time for R&D by about the. same proportion as .for
IH results in an estimate of 14 months for Chrysler. This
added to the 6 months needed to establish a test cell yields a
total of 20 months leadtime.
IH claimed that a total of 32 Months would be required to
implement this regulation. The first 12 months would be needed
to procure and install a heavy-duty dynamometer. As we have
discussed previously, this can now be reduced to 6 months
because of test procedure changes. IH then claimed that 6
months would be needed for R&D work. We will reduce this by 1
month because of the decreased R&D effort required due to less
family-systems, no testing of Class VII or VIII HDGs, and the
less stringent standard.
Furthermore, IH claimed that 14 months would be needed for
tooling once the new designs have been identified by the R&D
program. The question of tooling time deserves discussion.
The other three manufacturers did not specifically mention
tooling as a component of their leadtime estimations. This may
be because required changes to carburetors are expected to be
non-existent as far as tooling time is concerned or perhaps the
manufacturers perceived that while some tooling may be
necessary, . it would be so minimal that other leadtime items
would be the critical ones. Since we have reduced tne
manufacturers' leadtime estimates, we have investigated, tooling
time to be sure that it is still not a critical path for any
manufacturer.
Evaporative emission control is a very well understood and
simple technology. In fact, the majority of HDG carburetors
are virtually the same carburetors currently used on LDTs with
the difference being that the LDT carburetors are configured to
control evaporative emissions. We expect that for HDGs which
currently use these basic carburetors, very minimal tooling
(e.g., slightly larger bowl vent openings) will be reguireu.
These HDG carburetor modifications will almost certainly ue
made on existing production lines.
Even evaporative control system adaptations for those HUG
carburetors which do not have LDT counterparts will not have to
start from scratch. California has required evaporative
emission control of HDGs since 1978. Because the liug
manufacturers sell their vehicles in California, all uuG

-------
-80-
carburetors have designs for conversion to the evaporative
emission control configuration. These California control
systems are already controlling HDG evaporative emissions to a
level close to meeting the standards in this Final Rule,
especially since we have relaxed the standard for the larger
HDGs (Classes IV-VI) to 4.0 gpt. Thus, minimal, if any,
tooling will be required.
Staff experience indicates that minor carburetor changes
might need as few as six months to implement. A study by
Aerospace Corp.[l] indicates that normal tooling time required
to set-up an entire production line for a carburetor of known
design is about 12 1/2 months. This study estimated that such
tooling could be reduced to 11 1/2 months fairly easily. We
expect that any carburetor modifications necessitated by this
Final Rule will not involve the production of a new carburetor
but rather will be minor changes to the existing design of a
carburetor. This case is even simpler than the 11 1/2 month
case described in the Aerospace report because new production
lines will not have to be built. Instead, we foresee minor
changes to existing production lines based on extensive
experience with evaporative emission control of LDTs and
California HDGs. Therefore, we estimate that 10 months is a
liberal allowance for any tooling that this Final Rule may
necessitate.
IH's estimate of 14 months was based on the proposed
standard of 3.0 gpt for all HDGs. This Final Rule contains a
standard of 4.0 gpt for HDGs with GVWs greater than 14,GOD
lbs. All of IH's HDGs are greater than 14,000 lbs. GVW and,
therefore, will be subject to the 4.0 gpt standard. The less
stringent standard should allow IH to complete any required
tooling in 10 months. In any case, IH's required leadtime
estimate is 21 months.
The total leadtime estimates for the other manufacturers
are also well beyond the 10 months estimated for tooling.
Thus, we believe that even if these manufacturers retired
tooling, there exists adequate leadtime for it.
We have adjusted the manufacturers' estimates of required
leadtimes to 21 months for GM, 18 months for Ford, 21 months
for IH and 20 raonths for Chrysler because of changes to the
NPRM. Thus, based upon manufacturers'estimates, this
regulation can be implemented with the start of the 1985 muael
year for all manufacturers. Our independent analysis of t'.ie
required leadtime, which assumes that GM constructs a building
and that no manufacturer currently has test equipment,
indicates that 6 months would be required for test equij;.;ent
procurement and installation, 6 months would be needed for -au'o
(of which 50 percent can occur simultaneously with tujt
equipment installation), and 10 months would be required i£ any
tooling changes are necessary. Thus, the total leadtime ij

-------
-81-
about 19 months. GM's building construction would also fit
within this time-frame. Back-calculation of 19 months from
September 1984 gives a publication date as late as February
1983. Therefore, we conclude that if this Final Rule is
promulgated by February, 1983, then its implementation should
be the start of the 1985 model year.
The above analysis shows that implementation of this Final
Rule with the start of the 1985 model year will allow adequate
leadtime for all manufacturers. In fact, we expect that the
1985 model year implementation date will provide several months
of safety margin. This leadtime safety margin will reduce the
burden of this regulation even further from the already reduced
levels that were accomplished by changes to the level of the
standard, the certification procedure and the test procedure.
Extra leadtime will allow better planning for greater
efficiencies and will stretch out the financial commitment for
a better cash flow.
Recommendations
We recommend that the implementation date of this
regulation be changed from the proposed date (1933 MY) to the
start of the 1985 model year.

-------
-82-
References
1. "Assessment	of Domestic	Automotive Industry
Production Leadtime of	1975/1976 Model	Year: Volume II
Technical Discussion	Final Report,"	EPA-460/3-74-026^b,
December, 1972.

-------
-83-
E. Issue: Costs
Summary of Issue
In developing the. NPRM (45 F.R. 28S22) we estimated the
costs of the proposed regulation. These estimations are
detailed in Chapter V of the "Regulatory Analysis for the
Proposed Evaporative Emission Regulation for Reavy-Duty
Vehicles" (in the public docket). In that document we broke
down the expected costs to the manufacturers into four main
categories: 1) industry R&D costs, 2) industry investment
costs, 3) industry certification costs, and 4) control system
components costs (hardware).
For the proposal we projected that R&D costs would be
minimal since the technology for controlling evaporative
emissions is well known. Light-duty vehicle (LDV) and
light-duty truck (LDT) evaporative emissions have been
controlled for many years and the control of HDG evaporative
emissions was expected to closely parallel that of LDVs and
LDTs. We estimated that any necessary R&D costs would be fully-
accounted for by allowing a 100% profit markup from
manufacturing cost to retail price equivalent for the
hardware.
Industry investment costs included testing equipment such
as chassis dynamometers, SHEDs and durability testing equipment
in addition to a fair rate of return for the facility space
that would need to be allocated to HDG evaporative emission
testing. Total industry investment costs were estimated to be
$5.20H (discounted at 10 percent to 1983). When this total
cost was amortized over the number of new HDGs expected to be
produced from 1983 through 1S87 inclusive, the per vehicle
price increase due to industry investment costs was calculated
to be $2.91.
The proposal would have required each manufacturer to test
its HDGs and submit the data to EPA in order to receive a
certificate of conformity. This "certification testing" was
expected to be a major cost and was, therefore, estimated as a
major cost category. Components of certification testing costs
were expected to include personnel costs for control system
installation, triplicate testing and durability testing. Also
the cost of using the HDG during certification was induced.
These costs were estimated to be $2.37M for the five year
period of 1983 MY through 1987 MY. Amortized over the
projected HDG production during these five years this total
cost was calculateo to be equal to $1.33 per HDG.
The fourth and largest area of cost was that estimated Lor
the control system hardware. We estimated the typical control
system would consist of two charcoal canisters ($8.00 each),
extra hoses and tubing ($1.00), a carburetor fuel bowl vent

-------
-84-
($5.00)/ carburetor shaft seals ($1.00), a charcoal bed in the
air cleaner ($5.00), purge air intake from air cleaner ($.50),
a threaded fuel cap ($.50) and a liquid-vapor
separator/roll-over value ($1.00). These costs total $30 per
HDG.
Summation of the per vehicle costs for industry R&D,
industry investment, industry certification and control system
components yields, an estimated "sticker price" increase of
$34.20 per HDG.
Summary of Comments
Cost comments were received from all four of the primary
manufacturers (GM, Ford, Chrysler and.IH). Comments were also
received from two trade associations: the Truck Body and
Equipment Association (TBEA) and the National Automobile
Dealers Association (NADA). The primary manufacturers'
comments were general in nature with little detail in support
of their final numbers. This discussion will summarize each
manufacturer's comments separately and then turn to the trade
associations' comments.
GM categorized its expected costs into three areas: 1)
facilities and test equipment, 2) construction of data
vehicles, and 3) hardware costs. Its estimate for facilities
totalled $5.1M. The main component ($3.SM) of this cost was
the 26,000 ft^ that GM estimated it would need for 2 test
sites and 10 soak spaces. GM also included $0.5M for site work
and $0.7M as a contingency fund. GM provided no further
details. For test equipment, GM estimated a total of $1.4M
would be needed. The major items included: two dynamometers
(at $375K each), two computers (at $75K each), a canister
equilibrator (at $63K), two SHEDs (at $50K each) and a vehicle
scale (at $35K). Other equipment included two emission
benches, two balances, calibration weights, a driver's aid, a
power washer, six fuel carts, two vehicle movers and
miscellaneous parts and equipment. Again, no discussion or
other details were provided. GM's estimate for facilities ana
test equipment, therefore, totalled $6.5M.
GK's second area of cost estimates was the construction c£
data vehicles. GM estimated that the proposed regulation would
cause its HDG product line to be divided into 11 evaporative
emission families. GM then estimated that EPA would require it
to build and test three certification vehicles for seven ot
these families and two certification vehicles for the other
four. Thus, the number of certification builds was estimates
at 29. GM also estimated that it would cost between $20K anc.
$60K to buila each vehicle and run the certification tests.
GM's total estimated certification fleet cost was $1.27M ci;
which $C.164M was estimated to be recoverable due to salvage so

-------
-85-
that GM's final estimate for the construction of data vehicles
equaled $1.11K.
The last area of costs estimated by GM was that of
required hardware. GM stated that for HDGs "which are designed
and manufactured by our GMC Truck and Coach Division (generally
in the above 12,000 GVW range) we would require new designs and
hardware necessary to comply with the NPRM procedure and a A
g/1 standard ...the estimated hardware costs in. 1980 dollars
are listed below."
GM did not provide any details as to why these costs	are
estimated so high while its 1SS0 MY prices for California	KDG
evaporative emission control systems were priced at $122.66	for
single tank trucks and $161.40 for dual tank trucks.
Ford's comments were also very general. Ford claimed that
the proposal would cause it to incur investment expenditures
(including facility, tooling, launch and engineering costs) of
$12M (1980 dollars). This total consists primarily of test
facility requirements of $7M and engineering development costs
(including development testing). Ford did not reveal what
portion of the $12M total was represented by engineering
development costs nor did it describe the new
facilities/testing equipment that it claimed it would need.
Ford also stated that its projected cost ($12M) translated
into a first-year retail price equivalent (RPE) cost increase
of $60 per HDG vehicle (1S80 dollars). Ford claimed that if
HDGs with GVWs greater than 14,000 lbs were exempted from the
regulation then the RPE increase would be reduced to $40 per
HDG vehicle. Apparently this estimated reduction results from
an investment savings of $7M which would not have to be spent
on new facilities although Ford's comments are not clear on
this point.
The only direct mention of hardware costs in Ford's
comments was a general statement that the total investment cost
of $14M included come tooling. Tooling can be a major portion
of hardware costs. Ford's estimate of a $60 RPE increase must
include hardware costs since Ford compares that RPE increase
directly to the RPE increase of $34 that we estimated for the
proposal.
Chrysler's comments on expected costs were divided into
two areas: 1) hardware, ana 2) facility costs. Chrysler
stated that the proposed regulation would require a trap door
Vehicles with Engine
Single tank 292-L6 and 350-V8
Dual tank 366-V8 ana 427-V8
Estimated Cost
to Consumers
$250.00
$300.00

-------
-86-
air cleaner with dual snorkel, a second air pump (with brackets
and diverter value) for purging, three 5-port' canisters, a
sulfonated plastic gas tank and the tank vent system with
tubing. Chrysler did not give an estimated cost for each of
these components but rather gave a total cost of approximately
$150 (retail) for a HDG with a 45 gallon fuel tank while for a
HDG with a 36 gallon tank these hardware costs were estimated
to be $110. This is the extent of Chrysler's hardware cost
comments.
Chrysler listed a number of items and their estimated
costs under the general category of "Facility Costs". The most
expensive of these items was the square footage estimated for
fueling, soaking vehicles and general storage. Chrysler did
not state how much area was involved but did estimate the cost
to be $0.S3M. Chrysler also stated it would need a SHED (at
$180,000), a 200 hp dynamometer (at $.110,000), "analytical
facilities and test equipment" (at $85,000) ana contingency
funds of $130,000. Thus, Chrysler's total facility cost
estimate was $1.43 5M.
IH's comments
manufacturers. IK
dollars):
were the least
. presented the
detailed of the
following cost
four primary
chart (1960
Engines
Capital & Product Costs
Engineering ^Warranty
Total
Two-Barrel Engines Four-Bar rel
$53
$27
$80
$119
$ 27
$146
IH then stated, "The substantial increase in the cost of the
four-barrel configuration is due to the complexity of the
governed - dual float bowl four-barrel carburetor and its
associated tooling, plus low volume over which costs can be
spread." This concluded IH's comments on costs.
The comments from the two trade associations (TBEA and
NADA) were basically the same. They claimed that secondary
manufacturers could not afford to test HDGs according to the
proposed full-SHED test procedure. Therefore, any Final Rule
eventually promulgated must not require testing by secondary
manufacturer s.
Analysis of Comments
The cost estimates submitted by the four primary
manufacturers were, £or the most part, of a ver^ general nature
which restricted their usefulness for this rulemaking. Many of
the estimates lacktc sufficient detail to allow a substantive
analysis. For example, when a manufacturer merely states that
hardware costs will be $250 per vehicle or that investment

-------
-87-
costs will total $12M without including a breakdown of such
gross figures, there is little information for us to use in our
determination of the appropriateness of the estimate. These
broad estimates are further diminished in value when they
differ from our's and/or independent sources by large amounts
(5-10 times) such as some of the manufacturers' estimates do.
For these reasons, and because the Final Rule includes changes
to the proposal which were specifically developed to reduce
costs, the manufacturers' cost estimates were not as helpful as
they might have been.
The following discussion will examine each manufacturer's
cost estimates and will point out areas where those estimates
differ significantly from ours. Furthermore, the impact of the
changes to the proposal on the cost estimates will be
examined. This presentation will attempt to clarify the
comments and highlight those comments that were useful in
developing the Economic Impact chapter of the "Regulatory
Support Document." That chapter presents a detailed analysis
of the expected costs of this Final Rule. The "Regulatory
Support Document" can be found in the public docket
(#OMSAPC-79-l) for this rulemaking.
GM claimed it would need an additional 26 ,000 ft2 of
facility space because of the proposed regulation. This space
included 10 soak spaces and 2 test sites. If we assume a
typical soak space/test site is 14' wide by 40' long, then the
twelve such spaces claimed as necessary would sum to only 6,700
ft2. If the facility was laid out such that there was two
rows of six sites with 40' between the rows for maneuvering the
HDGs, the total space required is only 10 ,000 ft^. Even with
an allowance for fueling and storage areas GM's total facility
space requirement under the proposed regulation should not have
been more than 13,000 ft^. Thus, GM appears to have
overestimated the building cost by at least a factor of two
($ 2 .5 M) .
The changes to the proposal included in the Final Rule
significantly affect the amount of facility space needed by
GM. Under the proposal GM estimated that it would be required
to certify 29 family-system combinations. Under the Final Rule
we have estimated that GM will have only 8 family-system
combinations because of the provision for "worst case" control
system development and the redefinition of "evaporative
emission family." This reduction in the number c.f
family-systems will substantially reduce the size of the
testing facility because the number of vehicle control systems
to be developed will be reduced by about 70 percent. Thus,
GM's estimate of 20 , 000 ft2 ($5.1M) is out of proportion to
the requirements of the Final Rule.
GM estimated that new test equipment would cost $1.4to. in
its estimates GM assumed it would need two test sites. Due to

-------
-88-
changes incorporated in the Final Rule, we estimate GM will
need only one test site. Also, changes in the test procedure
will allow upgrading of existing light-duty dynamometers at a
cost of $25K instead of requiring the purchase of new
heavy-duty dynamometers at a cost (GM's estimate) of $375K
each. GM's estimate for new test equipment is thus reduced by
at least 2/3's.
GM also estimated that certification vehicle builds would
cost S1.11M per year under the proposed regulation. This
estimate included 29 family-systems at an average cost of
$38.3K per vehicle. The already mentioned changes to the
definition of "evaporative emission family" is expected to
reduce GM's family-systems to 8. Additionally, the Final Rule
does not require any certification, testing at all. This
provision will further reduce these costs since GM will not
have to build final development vehicles for those control
systems where there is a considerable safety margin of control.
The final area of GM's cost comments concerned the
hardware needed for the control systems. GM estimated that
hardware costs for HDGs above .12 ,000 lbs. GVW with a single
fuel tank would be $250.00 per HDG ($300.00 for vehicles with
dual fuel tanks) . The only support for these numbers was a
statement that 1S80 MY prices for California HDG evaporative
control systems were $122.66 (single tank) and $161.40 (dual
tank) and that the control system required by this regulation
would be "more sophisticated and, therefore, possibly more
costly than the 1980 MY system (California)."
First, it appears somewhat contradictory to state that the
new control systems will "possibly" be more costly and then to
estimate the new cost at double that of the old system.
Second, our estimates of hardware costs, which include liberal
profit markups and were developed by Exxon,[1] totalled
$38.25. Third, Ford estimated that the first-year retail price
equivalent (RPE) cost increase for the proposed regulation
would be about $60 per HDG which includes $.12M for investment
requirements. The cost of just the control system hardware
would be substantially less than the total of $60 per HDG.
Ford, also stated that if HDGs with GVVvs greater than 14,000
lbs. were exempted from this rulemaking, then the RPE increase
would only be $40 per HDG since control!ing the heavier HDGs is
more costly. This Final Rule contains a split standard where
HDGs with GVNs greater than 14, 000 lbs. are subject to a 4.C
gpt standard instead of the proposed 3.0 gpt standard. The
split standard was developed so that the heavier HDGs woula not
require substantially more R&D effort than the lighter HDGs.
This will decrease the overall RPE increase for HDGs. Ford's
hardware cost estimate, therefore, agrees quite well with cu
while GM's hardware costs appear to be considerably
overestimated.
L O

-------
-89-
Fora's cost comments were, like GM's, too general to be of
use to the degree we would have expected. As discussed above,
Ford stated that the RPE increase under the proposed rule would
be about $60 per HDG. However, Ford did not separate this
figure into hardware vs. investment costs and did not itemize
required hardware at all. Ford did state that investment costs
were expected to be $12M ana included test facility
requirements of $7M with tooling, launch and engineering
expenditures making up the rest. Ford did not give any detail
as to how the above investment expenses were broken down. We
have concluded that Ford would not need to spend more on
facilities than GM and GM's required facility expense under the
proposed regulation would not have exceeded $3M. Thus, even
without those changes to the proposal which significantly
reduce the required facility and test equipment expenses, we
have determined that Ford's facility estimate is about 233
percent larger than our estimate. Ford's required facility,
like GM's, is substantially reduced due to the changes in the
vehicle classification scheme, changes in the certification
procedure and changes in the test procedure. Without a
detailed breakdown of its estimate, it is hard to know what
Ford's new facility could have included that would have cost
$7M. Ford's estimate cf tooling, launch and engineering
expenditures suffer from the same lack of detail.
Chrysler's hardware cost comments were the most detailed
of the four manufacturers. Chrysler listed five items which it
claimed would be needed to control evaporative emissions from a
HDG with a 45 gallon tank. Chrysler did not give individual
pr ice estimates for each item but rather gave only a total
price for all five of $150.00. Apparently the difference
between our cost estimate for hardware and Chrysler's is that
Chrysler claimed the proposed regulation would require it to
equip HDGs with a second air pump for purging and a sulfonated
plastic gas tank. The other items listed are common to our
list and we will assume that their prices are approximately the
same.
Chrysler does not discuss the need for a second air pump
except to state that it is needed "for purging the evaporative
emission system." Our analysis indicates that a second air
pump will not be necessary. Ford did extensive testing of HDGs
to determine the technical feasibility of the proposed
standard. The results of Ford's testing showed that the HDG
control system coulo be adequately purged over the proposed
driving cycle with existing equipment. In fact, Ford's
California HDG evaporative emission control systems actually
met this standard in many cases. Also, IH did some testing and
did not indicate a need for an extra air pump to increase
purging. We believe that once Chrysler begins to test its HDGs
it will agree that the IiDG driving cycle provides adequate
purging opportunity without the addition of an extra air pump.

-------
-90-
Chrysler's claim that it would need a "sulfonated plastic
gas tank" as a result of this regulation is not discussed at
all in Chrysler's comments. We have no idea why this
regulation would necessitate the use of a sulfonated plastic
gas tank. Perhaps Chrysler is already using plastic gas tanks
and this regulation would require the sulfonation of these
tanks to reduce HC vapor permeability. However, sulfonation
would be a minimal expense.
Chrysler's facility costs estimates are approximately the
same as outs. However, due to changes to the proposal Chrysler
will not need a $11GK dynamometer. Instead, a $25K retrofit
kit will be needed. Also, Chrysler estimated that a new SHED
with associated analytical facilities and test eqipment would
cost $265K. Our estimate which is based on discussions with
SHED manufacturers, is $141K.
Finally, Ill's cost estimations were too general to be of
much use in evaluating the costs of the final rulemaking. It
should be noted, however, that because we have changed the
level of the standard that IH must comply with from 3.0 gpt to
4.0 gpt (IK does not produce any HDGs with GVWs less than
14,001 lbs.), the extra complexity of controlling the
governed-dual float bowl four-barrel carburetor which lb
claimed would add a substantial cost to the RPE increase will
be significantly diminished. In fact, IH is very close to
meeting the' 4.0 gpt standard now with its California HDG
evaporative emission control systems as shown by its testing
effort for this rulemaking. A discussion of IH's testing
effort can be found in "Issue C: Technical Feasibility" which
is a part of this document.
The above discussion has shown that the manufacturer's
cost comments were limited and of a general nature which
restricted their usefulness in the estimation of the cost of
this Final Rule. Our analysis of the economic impact of this
Final Rule can be found in Chapter V of the Regulatory Support
Document (in the docket).

-------
-Sl-
References
1. "Investigation and Assessment of Light-Duty Vehicle
Evaporative Emission Sources and Control," P.J. Clarke, Exxon
Research and Engineering Company, EPA Report No.
EPA-460/3-76-014, June 1S76.

-------
-92-
F. Issue: Test Procedure
Summary of the Issue
The proposal for control of HDG evaporative emissions
which was published on April 30, 1980 included a testing
procedure (Subpart I'i) . Test procedures are needed to assure
EPA that r.vanuf acturer s are attaininy the required level of
control as well as to identify the in-use levels of control for
air quality modeling and planning. The test procedure, as
proposed, is similar to that used for light-duty vehicles and
liyht-duty trucks. However, changes to the light-duty test
proceuure have been made where necessary to accomodate the
differences between light-duty and heavy-duty vehicles.
The test procedure requires that a HDG be placed in an
airtight enclosure (Seal Housing for Evaporative Determination
(SHED)) and its evaporative emissions measured. The test
procedure can be divided into four major parts: 1)
preconditioning, 2) diurnal, 3) driving cycle, and 4) hot
soak. The purpose of the first part, preconditioning, is to
stabilise and appropriately load the carbon canisters in
preparation for the rest of the test procedure.
Virgin-activated charcoal will adsorb more HC vapors than will
activated charcoal which has been through many cycles of HC
vapor adsorbtion/desorbtion. Therefore, the test procedure
requires that charcoal canisters undergo a number of HC vapor
adsorbtion/desorbtion cycles so they will simulate the
canisters on in-use vehicles. Also, this preconditioning phase
will load the canisters so that the canisters enter the next
part of the test procedure with some HC vapors in them as they
would under real world conditions.
The second part of the test procedure attempts to simulate
the effect of a summer day on evaporative emissions. As the
heat of the day increases so does the temperature of the fuel
in the tanks. The resultant expansion of air in the fuel tanks
causes HC vapors to be forced out of the tank. These vapors
must be trapped and disposed of properly rather than being
emitted to the atmosphere. In the test procedure the fuel
tanks are filled tc 40 percent of capacity with chilled (less
than bQ^F) fuel. The HDG is then pushed into the SHED and the
enclosure is sealed. Heat is then applied to the fuel tanks so
as tc raise the fuel temperature from 60° F to 84° F over a
1-hour period. The HC vapors that escape the control system
during this hour dee measured (in grams) and constitute the
result of this diurnal part of the test.
The vehicle is then either pushed or driven to the
dynamometer where it is operated over the driving cycle given
in Appendix 1(d). mis serves two major functions. First, the
engine and engine compartment are heated in preparation foe trie
hot-soak part of the test procedure. Second, this engine

-------
-93-
operation will purge the carbon canisters of KC vapors thereby
making room for more HC vapors generated during the next part
of the test.
The final part of the test procedure is called the hot
soak. After operating the vehicle over the driving cycle, it
is again placed into the SHED. The SHED doors are sealed and
the concentration of HC in it is monitored for one hour. The
heat from the engine block arid engine compar tment will raise
the temperature of the vapor space of the carburetor fuel
bowl. The resultant expanding vapor will seek to escape
through leaks in the induction system (e.g., throttle shaft,
accelerator linkage, air cleaner). An evaporative control
system will attempt to route these vapors to a carbon
canister. The HC vapors which escape the control system are
measured after one hour, and the result, is added to the result
of the diurnal to arrive at the total test result which is
expressed in grams HC per test (gpt).
Summary of the Comments
GM, Ford, Chrysler, and tWMA were the only commeriters on
this issue. The comments received can be divided into major
and minor categories. Major comments are those which require
discussion and analysis in some depth. Minor comments are
those which are simple to resolve and require very little
analysis. Minor comments include grammatical errors,
typographical errors, and obvious omissions. These minor
comments are treated separately as an appendix to this issue.
This section will briefly summarize the major comments.
Further explanation and detail will be included in the next
section, Analysis of the Comments.
GM, Ford, and MVMA commented that the proposed formula for
determining the dynamometer loading is incorrect. GM claimed
that many HDGs have the same physical body shells as LDTs and,
therefore, must have the same coefficient of aerodynamic drag.
However, this coefficient is .50 for light-duty vans and .53
for light-duty pickup trucks while it would be .67 for all HDGs
under the proposed test procedure. Thus, the dynamometer
loading would be higher than it should be for HDGs with van or
pickup body styles. Furthermore, GM claims that the hug
dynamometer tire friction factor is too- small and also
questions why a "coastdown" procedure is not allowed.
Ford presented coastdown data from a C-700 vehicle wiiich
showed that the proposed dynamometer loading formula results in
a 36 percent higher dynamometer load than actual road load
requirements would dictate. Ford also presented an alternative
method for deteruining the dynamometer setting but cautioner
that more data was needed to verify the experimentally derived
formulas.

-------
-94-
All four commenters stated that the proposed test weight
of 70 percent GVW is too much. They pointed to a study
conducted by Olson Laboratories, Inc.,[l] an FEA document,[2]
and a letter dated November 6, 1975, from former EPA Director
of Technology Mr. John DeKany to Mr. Duffing of 1WMA [Appendix
1] as evidence that the typical HDG does not weigh 70 percent
of GVW. GM also stated that, at first glance, a document by
the FHA[3] appears to support the 70 percent of GVW vehicle
test weight requirement. GM then listed four reasons why this
document is not representative of HDGs and concluded that it
should not be used in the determination of the proper test
we ight.
GM, .MVMA, and Chrysler claimed that the proposed
requirement that ilDGs be run over the driving cycle with hood
closed is unreasonable. They stated that the light-duty
evaporative test procedure calls for testing with the hood open
in order to prevent overheating. Since HDG testing presents a
more serious cooling problem, EPA should allow the hood to be
open and should allow the manufacturers to employ cooling fan
capacities greater than 10,000 CFM (§86.1235-84). They claimed
that closing the hood for HDG testing would cause unrealistic
and excessively high levels of evaporative emissions to be
formed.
Another major area of comment was the driving cycle. GM
stated that both the HD engine cycle and the proposed hd
chassis cycle are derived from the CAPE-21 data base. Since
the engine cycle is 107 seconds longer than the chassis cycle,
it is not possible to claim that both cycles are
representative. GM, Ford, and Chrysler were concerned that the
shorter cycle would allow less purge time and thus make control
more difficult. GM requested that consistency between the two
test procedures be established and that the record be kept open
on this issue until MVMA completes its study of EPA's
derivation of the cycles from the CAPE-21 data. Chrysler
agreed that EPA should wait for the MVMA study to be completed
and also claimed tnat the chassis cycle was unrepresentative of
its HDG product line because Chrysler HDGs are all in the lower
weight classes while the CAPE-21 data included many HDGs in the
upper weight class.
GM and Chrysler disagreed with EPA's proposed 24° F ruel
temperature rise for the diurnal part of the test.. GM cited
two documents! 4,5 ] as evidence that 24° F was too high. l>M
concluded from the EPA document that a typical uaiiy
temperature excursion was 20° F, not 24° F, and tnat ti.e
average fuel tank is filled to 59 percent capacity instead oi;
the 40 percent capacity proposed. GM stated that the
time/temperature document does not support the HPHM tc-yt
procedure either because the daily temperatures were auuve
normal and the vehicles studied were passenger cars. ij:-i
recommends a 15° F fuel temperature rise with a 60 percent lu^L
fill.

-------
-95-
Chrysler stated that the 24° F fuel temperature rise is
unrepresentative of its HDGs because EPA based the 24° F rise
on HDGs with saddle tanks (i.e., those fuel tanks which would
be directly exposed to the sun's rays) and Chrysler does not
use such fuel tanks on its trucks. Also, Chrysler claimed that
the driving cycle was based on a delivery type of schedule yet
the diurnal heat build assumes a commuter type of schedule.
Chrysler objected to this inconsistency. Chrysler recommended
a 15° F fuel temperature rise and presented a heat flow
calculation to substantiate it.
Analysis of the Comments
Test Weight
We agree with the commenters that 70 percent of GVW is not
the average weight of a typically loaded HDG. However, at the
time this test procedure was developed, it was not the intent
of the Ayency to address the "average" HDG. Instead, we wanted
to ensure that the majority of HDGs would comply with the
standard. We believe that an adjustment to the proposed test
weight is appropriate because of reasons in addition to the
manufacturers' comments. We will begin this discussion with a
brief analysis of the manufacturers' comments and then turn to
the above mentioned reasons.
The letter from Mr. DeKany of EPA to Mr. Duffing of MvHA
(Appendix I), which all four commenters cited as evidence
against the 70 percent factor, gives a "mean-weighted payload"
of about 500 lbs. for trucks in the 6 ,000 to 9,000 lbs. GVW
range. If we assume that a HDG with a GVW of 8,600 lbs. has a
curb weight of 4,092 as stated by GM, then it might be
considered that the test weight should be 4,092 plus 500 lbs.
or 53 percent of GVW. However, the "mean-weighted payloau"
figure is an average of all trucks whether loaded or not.
The DeKany letter also gives a number called the
"mean-loaded payload." This number is an average for only
those trucks which had a payload and is thus more pertinent to
a strategy of controlling the majority HDGs rather than just
the "average" HDG. The "mean-loaded payload" for the 8,600 lb.
GVW truck discussed above is estimated to be about 78U lbs.
from the DeKany letter. The test weight for this truck would
be 4,092 lbs. (GH's curb weight estimation) plus 780 lbs. ( the
"mean-loaded payload") for a total of 4,872 lbs. or 57 percent
of GVW.
The above calculations assume that the curb weight of
trucks have significantly been reduced (as GM claimed) since
the time that the uata on which the DeKany letter is baseu was
collected (1973). If the curb weights given in the DeKany
letter were useu in conjunction with the "mean-weiy liteu
payload" (500 lbs), then the test weight for a HDG with a GVW

-------
-96-
of 8,000 lbs. would be 6,600 lbs. (6,100 lbs., curb weight plus
500 lbs mean-weighted payload" ) or 33 percent of GVW. For a
HDG with a GVW of 9,000 lbs. the same calculation yields a test
weight of 80 percent of GVW. While we have no reason to uoubt
GM's claim of reduced curb weights relative to GVW, the DeKany
letter alone would certainly not support lowering of the
proposed test weight.
The data used in the DeKany letter was 1973 FHA weigh
station survey data. This same source (but a different year's
data (1977)) is later attacked by GM as nonrepresentative. Gh
claimed it is not truly representative since only interstate
highway vehicles are in the sample size ana the curb weights
are in excess of late MY (1980) vehicles. Both points are well
taken and apply to the DeKany letter as well as the 1977 FHA
data. Due to the push for greater fuel economy, curb weights
have indeed dropped substantially and GM's estimates seem
reasonable. Also, we agree that local as well as interstate
traffic should be considered in setting the test weight
requirement. The FHA data only includes interstate traffic,
but when combined with GM's curb weight estimates, it indicates
that the test weight should be in the range of 50 to 60 percent
of GVW.
A better source for determining the test weight is the
Federal Energy Administration (FEA) study because it includes
both local and intercity traffic. When the average vehicle
loads of this study are combined with GM's curb weight values
for the appropriate GVW classes, the test weights are
calculated to be about 50 percent of GVW for all HDGs less than
26,000 lbs. GVW. (HDGs having GVWs greater than 26,000 lbs.
will not be tested but will be subject to an engineering
evaluation instead.)
The data sources also indicate that the loaded vehicle
weight as a percent of GVW varies with GVW. For maximum real
world simulation some relatively small increment of GVW would
need to be selected and a different percentage applied to eacn
GVW increment to derive the test weight. This would be a
sizeable task and the resultant improvement in the test
procedure would be minimal because a 10 or 20 percent increase
or decrease in test weight is not going to affect the purge
rate or engine compartment temperature in a major w>;. We
conclude that the proposed method of using one percentage for
all GVWs will be appropriate for all sizes of hsGs.
After reviewing that data available, we have determineo
that while the proposed test v.^ight of 70 percent of GVW v,ould
be most appropriate i. f a "worst case" strategy of control were
pursued, a test weignt of 50 percent of GVW is more appropriate
if an "average ca^e" strategy is pursued. On the other hand,
because ol tne jlj...;ilcu toiu that the test weight plays in the
test procedure, it is not likely c-.a.t the difference between

-------
-97-
these two choices will substantially affect test results.
Therefore, since we wish to reduce the burden of this
regulation to the extent possible and since we do not believe
the change in test weight will significantly affect air
quality, we recommend that the test weight be changed to 50
percent of GVW.
Dynamometer Load Formula
GM and Ford commented in some detail on this topic. While
the main thrust of their comments were the same (i.e., EPA's
proposed dynamometer load formula produces horsepower settings
that are too high), they attack the issue from different
viewpoints.
GM claimed that EPA's frontal area coefficient of .67 is
incorrect. GH stated that the formula for LDTs uses a frontal
area coefficient of .50 for vans and .53 for pickup trucks.
Since some HDGs in the lower weight classes have pickup or van
body styles, it stands to reason that these vehicles shouia
have frontal area coefficients the same as for LDTs. However,
under the E1DG test procedure, the .67 frontal area coefficient
would have to be used, thus resulting in a higher dynamometer
horsepower setting than if the LDT coefficient had been used.
We agree with GM that the .67 frontal area coefficient is
not accurate for all HDGs. It is too high for the smaller HDGs
and it is too low for the large HDGs. GM suggested that EPA
propose multiple factors to cover each body shape found in the
heavy-duty truck cateyory. We have determined that the cost
(e.g., testing, person-hours, etc.) to develop such a set of
factors is greater than the relatively small benefit that would
be gained. A different situation exists for LDTs in that two
main body shapes (van and pickup) account for the large
majority of LDTs sold. This situation does not occur with
HDGs. HDG's body shapes include buses, tractor/trailer
combinations, pickups, large delivery vans, garbage trucks,
small vans, and stake-body trucks to name a few. EPA would
need to expend a significant amount of resources to determine
the proper frontal area coefficient for all HDG body types.
The proposed coefficient of .67 was intended to be an average
which we believe to be representative of all liDGs.
While we agree that the .67 coefficient is high for some
HDGs, the purpose of the coefficient must be considered in
determining if it is too high. In LDT certification, tne
frontal area coefficient is used to determine the dynamometer
horsepower setting. This is also true for HDGs. However,
there is a major difference between the two classes in t.ie
purpose of running the vehicle on the dynamometer. In ll>T
certification, the truck is run on the dynamometer and exhaust
emissions are sampled. Almost as a by-product, the vehicle's
engine compartment is warmed and the carbon canisters are

-------
-98-
purged in preparation for the hot soak part of the evaporative
emission test. The affect of different dynamometer horsepower
settings is much greater on exhaust emissions than on
evaporative emissions. As will be discussed below, a
relatively large difference in horsepower settings produces a
small difference in total evaporative emissions. It is
understandable that LDT frontal area coefficients should be
relatively more accurate because they are used in exhaust
emission testing as well as the warm-up cycle for evaporative
emiss ions.
We believe that retaining the .67 frontal area
coefficient, while being slightly more stringent for some of
the lower weight HDGs, will not result in any unrepresentative
dynamometer horsepower settings. Some settings may be higher
than the "average" truck for that body shape category but still
will not approach the "worst case" situation. In this sense,
the .67 coefficient is lenient. We also point out that the
dynamometer horsepower settings for HDGs as determined by this
test procedure are already being lowered as a result of the
change in the test weight formula. We proposed the test weight
to be 70 percent of GVW. As discussed in the previous section,
we now recommend that the test weight in this Final Rule be 51)
percent of GVW. This results in a lowering of the horsepower
setting of about 10 percent.
Ford did not specifically assail the .67 frontal area
coefficient. Instead, Ford conducted a coastdown testing
program with a C-700 HDG (27 ,500 lbs. GVW) to determine the
actual track road load. This vehicle was selected for testing
because it exhibited the most difficulty in meeting the maximum
speed requirements of the driving cycle proposed by EPA. With
the vehicle loaded to 70 percent of its GVW, Ford's coastdown
testing yielded a dynamometer setting of 55 hp. The EPA
formula would result in a dynamometer setting of 75 hp. Thus,
Ford's coastdown testing of one HDG resulted in a dynamometer
setting 27 percent lower than that derived from EPA's
dynamometer load formula.
Ford then presented an alternate method for determining
the dynamometer setting. However, Ford cautioned that this
alternate calculation uses experimentally-derived formulas
which are based on the data from only one vehicle. Ford stated
that further data on a variety of vehicles must be obtaineu to
verify the method.
While a 27 percent difference in dynamometer horsepower
setting may seem to be substantial, Ford's test data indicates
that it probably doesn't make much difference in evaporative
emissions testing. Ford tested two C-700 HDGs for evaporative
emissions using the proposed test procedure. The tests
pertinent to this discussion were those in which • the test
weight of the vehicles were held constant at 19,250 lbs. (70

-------
-100-
have the most trouble since such vehicles would be in WOT more
frequently than th lighter HDGs. Chrysler, which only produces
the lighter HDGs, should have no trouble adequately purging the
carbon canisters using this driving cycle.
Chrysler's other concern was that since it produces only
the lighter HDGs, the driving cycle might be unrepresentative.
It is true that the driving cycle is derived from a data base
which includes operational characteristics from light-,
medium-, and heavy-weight HDGs. However, we do not believe it
is reasonable to expect a different driving cycle for each
manufacturer depending on its particular product mix. The
Agency simply does not have the resources to undertake such a
task. We believe the driving cycle is representative of the
HDG class.
GM claimed that since this driving cycle has a duration of
1,060 seconds and the exhaust emission cycle (engine transient
test) has a duration of 1,167 seconds, one or the other must be
unrepresentative.
Both cycles are representative. The technique used to
generate all of the cycles was a random number generator with
appropriate filters and is known as the "Monte Carlo"
technique. The mean trip time as determined from the CAPE-21
data base was approximately 18-20 minutes. In generating
driving cycles from the CAPE-21 data base it was desirable to
keep the time near 18-20 minutes. However, it would be unusual
to have each cycle exactly the same length of time due to the
randomness of the "Monte Carlo" technique. In fact, because
the cycle would need to be forced to terminate, it would not be
randomly generated thus reducing the objectivity of the "Monte
Carlo" technique. The 1,060 seconds duration of the HDG
evaporative emission cycle equals 17.67 minutes. The diesel
engine transient test cycle is 1,199 seconds (i.e., 19.98
minutes), and the gasoline engine transient test cycle is 1,167
seconds (i.e., 19.45 minutes). While none of the three cycles
mentioned above are exactly the same duration, they are all
about 18-20 minutes long. They are all within the "window"
that was established for representative cycles. It should also
be repeated that Ford has already shown, through its HDG
evaporative emission testing program, that the driving cycle
does allow adequate time for purging. We assume that this is
why GM was concerned with the discrepancy between the duration
of the exhaust cycle and this evaporative cycle.
The representativeness of the CAPE-21 data base and the
appropriateness of the "Monte Carlo" technique were both
analysed in detail during the Final Rulemaking for the 19d4
heavy-duty engine exhaust regulations. The reader is referred
to the "Summary and Analysis of Comments" for that rulemaking
(in Public Docket if OMSAPC-78-4) for further information on tne
development of the cycles.

-------
-101-
Fuel Temperature Rise
Both Chrysler and GM claimed that a 15°F fuel temperature
rise would be more appropriate for HDGs than the proposed 24°f
fuel temperature rise. Chysler claimed that the driving cycle
was based on a delivery type of schedule yet the diurnal heat
build assumes a commuter type of schedule. Also, Chysler
presented a heat flow calculation which showed that given the
same ambient temperature and fuel tank configuration, the fuel
temperature in a large tank will increase more slowly than that
in a small tank. GM claimed that the two documents[4,5] we had
used to substantiate the 24°F fuel temperature rise instead
supported a 15°F fuel temperature rise.
The CAPE-21 data base included HDGs of many occupational
uses. The resultant driving cycle is representative of the
class of HDGs. We would agree that the majority of HDGs
instrumented for the CAPE-21 study were in some sort of
delivery mode. However, we do not agree with Chrysler that the
heat build assumes a commuter type of schedule. Since Chrysler
did not expand on its comment, it is not clear what it means by
a "commuter type of schedule." The 24°F heat build assumes
only that the vehicle is outside for the entire day. This is
certainly true foe the great majority of HDGs.
While we agree with Chrysler's general calculation, there
are some specific factors which it did not take into account.
First, there are many HDGs in the lighter weight categories
which have fuel tank capacities similar to light-duty trucks
and undergo the 24°F fuel temperature rise of the light-duty
test procedure. Fifty-three percent of all HDGs have GVWs of
14,000 lbs. or less. This means that while there are some HDGs
with fuel tank capacities of 100 or even 150 gallons, many HDGs
have fuel tanks in the 20 to 40 gallon range, which is similar
to LDTs. Second, for the most part HDGs with the large fuel
tank capacities have these fuel tanks located where they are
exposed directly to the sun. The fuel in these "saddle-type"
tanks would thus be expected to increase in temperature more
quickiy than the fuel in tanks which are located underneath the
vehicle and thus shaded.
Finally, Chrysler did not consider the fact that during a
daily ambient heating cycle the fuel temperature of the smaller
tank will cease to rise long before that of the larger tank.
This occurs because the ambient temperature rises until it
reaches a maximum and then decreases to its minimum.
Chrysler's calculation assumes a constant heating rate. The
fuel temperature of the smaller tank will lay behind the
ambient temperature and will continue to rise until the ambient
temperature has begun its descent and equals the fuel
temperature of the smaller tank. From the point where the
ambient temperature and the fuel temperature of the smaller
tank are equal (i.e., maximum small tank fuel temperature) to

-------
-102-
the point where the ambient temperature and the fuel
temperature of the larger tank are equal (i.e., maximum large
tank fuel temperature) the difference between the fuel
temperatures of the two tanks will shrink. Thus, the
difference between the fuel temperatures of the two tanks that
one might observe at the maximum ambient temperature can be
expected to decrease throughout mid-afternoon.
GM claimed that the two documents[4,5] indicated that the
fuel temperature rise during a typical day equals about 75
percent of the ambient temperature rise and that the typical
summertime ambient temperature rise is from 64°F to 84°F.
Thus, GM recommends the test procedure fuel temperature rise
should be 75 percent of 2Q°F which equals 15°F. We a^ree with
both of these findings, however, we do not agree with the
recommendation because we are concerned with the control of
evaporative emissions in most situations rather than just 50
percent of them. The days when higher than average ambient
temperature diurnals occur are generally the same days when the
amount of sunlight is most and wind is least. These factors
aggrevate the air pollution problem and, therefore, the release
of HC vapors should be doubly guarded against. We believe that
this test procedure should attempt to assure control of
evaporative emissions during these critical periods as well as
during the average periods.
The typical summertime ambient temperature diurnal to
which GM referred is exceeded many times. In fact, a 32°F
ambient diurnal is not at all uncommon as Figure 1 indicates.
A test procedure fuel temperature rise of 75 percent of such a
32°F ambient diurnal would yield 24°F. This is the fuel
temperature rise of the proposed test procedure and we
recommend its retention in the Final Rule.
Canister Preconditioning
GM claimed that the proposed canister preconditioning
requirements were excessive. We had proposed that "green" or
"virgin" activated carbon in the evaporative control system be
exposed to 100 adsorb/purge vapor cycles before the full-SHED
test would be run. Adsorb/purge vapor cycles are necessary to
stabilize the carbon's working capacity. The 100 cycles
proposed was estimated to be a sufficient number to assure
stabilization. Ninety of the 100 cycles could be done with a
bench-type procedure while the last ten cycles had to be done
while the carbon device was attached to the HDG in its
production configuration.
GM submitted data showing that equilibrium was reached
after approximately 20 adsorb/purge cycles. We have,
therefore, changed this Final Rule accordingly. The official
test procedure requires only 20 bench-type adsorb/purge vapoc
cycles plus ten cycles with the carbon device attached to tne

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
-103-
OF DAYS
0	10	20	30	40	50
<~
•
«~







o-<~<~<>• oo o-0^»-» >



~ ~~ ~'»•»<» <~-»~ » ~ <~ » ~










1 »»4



>»*»
Min
Max
Mean
s
3°F
45°F
20.9°F
»9»







Figure 1 Composite Distribution of Differential
Daily Temperatures - 5 Cities
(July 5. August)
Source: "Typical Vehicle Diurnal" U.S. EPA, October, 1976

-------
-104-
HDG. However, it should be noted that this aspect of the test
procedure is, in actuality, left to each individual
manufacturer's discretion. Each manufacturer is free to use
any testing method it chooses to assure compliance with the
standard and the only testing EPA might do would be on in-use
vehicles. Such vehicles will have canisters which are aged and
stabilized. The canister preconditioning aspect of this
official test procedure will thus not be necessary when and if
EPA ever does test HDGs for evaporative emissions.
Hood Open vs. Closed
MVMA, Chrysler and GM claimed that since the light-duty
test procedure requires the vehicle's hood be open during
operation over the driving cycle, this HDG test procedure
should at least leave that option up to the manufacturer. The
proposed HDG test procedure required that the vehicle's hood
remain closed. The commenters claimed that this was not
equitable with the light-duty procedure.
We acknowledge the fact that this HDG test procedure is
different than the light-duty test procedure. We do not,
however, find anything inequitable in this. HDGs' hoods are
certainly closed when they are in-use and leaving the hood
closed during the driving cycle therefore better simulates the
real world. The practive of opening the hood for light-duty
vehicles was established many years ago to aid cooling since
the fans available were, inadequate. We have doubled the
maximum fan size (from 5300 cfm to 10,600 cfrn) for this HDG
rulemaking. We believe these bigger fan capacities will
adequately cool the HDGs as they are driven over the cycle.
If, however, the manufacturer can show that during field
operation the vehicle receives additional cooling, and that
such additional cooling is needed to provide a representative
test, the fan capacity may be increased or additional fans used.
None of the commenters mentioned the increased maximum fan
capacity (as compared to light-duty). In a sense, this
increased maximum fan capacity is intended to offset the
reduction in cooling due to the requirement that the hood be
closed. None of the commenters claimed that 10 ,600 cfm would
not be enough to provide adequate cooling. In fact, Ford has
demonstrated the technical feasibility of the test procedure
and standard for a 27 ,500 lbs. GVW vehicle. We do not believe
that this test procedure must be identical to the li^ht-uuty
test procedure anu, therefore, recommend that the hood-closed
requirement be retaineo in the Final Rule.

-------
-105-
References
1.	"The Composit ion, Function and Travel Patterns of
Medium Duty Trucks," MVMA #L1751G-C1.21.
2.	"Trucking Activity and Fuel Consumption - 1973,
1980, 1985, and 1990," FEA/D-76-390, July 1976.
3.	"1977 Federal Highway Administration In-Use Truck
Weight Data."
4.	"Typical Vehicle Diurnal," Emission Control
Technology Division of the Office of Mobile Source Air
Pollution Control, EPA Evap 76-3, October 1976.
5.	"Time-Temperature Histories of Specified Fuel
Systems, Volume I," Coordinating Research Council, Inc.
(APE-5-68, October 1959.

-------
APPENDIX I
\\ry? ' ^ UNITED STATES ENViSGNW-lNTAL PROTECTION AGENCY
• • * ; " i G AN -C*. C3
Ncvesber 6, 1975
c?- cl
AlP iNO	»!
... _ vi«.	Du • t — •"* t
lot^r i^ia Mar.ufacturers Assoc.
"•2.. '.'<.- Jc.r.ter ,3uiicir.g
.we::::, Michigan -oICz
jci: X.r . Zu.iLing:
This is a response ro your request for information en light .duty
truck lading contained in your, letter of October 17, (attached) &?A
conducted an analysis of 1973 truck weigh station survey data available
froc :hc U.S. Dept. of Transportation, Federal Highway Administration
and thereby arrived at the figure of 500 lbs. average combined passenger
and car>;o Loading for light duty trucks. The table below cp—-¦¦•< -he
dat.i ui»*id in deriving this value.
TAIiUI-TRUCK WEIGHTS AND PAYLOADS (lbs)

Mean
Mean loaded
Mean .weighted
tPvW
Curb weight
payload
payload
6000
4695
905
5.13
7000
5280
770
470
SOUO
6100
734
456
9000
6671
809
534
The mean weighted payload value was derived, by nulcip Lying. the ^ean
lauded payload by the percent of trucks observed to be leaded and then
adding the 300 pound value currently associated with passenger ar.d fuel
loading for personal transportation vehicles. The value so detcminid
stays relatively constant over the GVW range under consideration for the
rcviicc light duty truck class anc (Therefore the approximate value of
500 lbs. vis chosen as representative.
Sincerely yours-,
C. Cray
C. Rossow

-------