EPA-AA-TEB-88-01
Formaldehyde Sampling From Automobile Exhaust: A Hardware Approach
William M. Pidgeon
July 7, 1988
Test and Evaluation Branch
Emission Control Technology Division
Office of Mobile Sources
Environmental Protection Agency
-------
EPA-AA-TEB-88-01
Formaldehyde Sampling From Automobile Exhaust: A Hardware Approach
William M. Pidgeon
July 7, 1988
Test and Evaluation Branch
Emission Control Technology Division
Office of Mobile Sources
Environmental Protection Agency
-------
Table of Contents
i. Introduction 3
2. Summary 3
3. The NPRM Required Sampling System 4
4. TEB's Formaldehyde Sampling System Design Considerations 4
4.1 A Basic System for Formaldehyde Sampling 4
4.2 The Proportional Flow Problem When Using a Heated Probe 5
4.3 The High Pressure Drop Problem 6
4.4 Variable Sample Flow Rate 8
5. The TEB Formaldehyde Sample System 8
5.1 Thermal Mass Flow Controller System Performance 11
5.2 Flow Controller and Sample Pump Selection Interactions 12
5.3 Flow Controller and Sample Pump Selection Decisions 16
5.4 Additional Details of TEB's System 17
6. List of Suppliers 19
Figure 3: Formaldehyde Sampling System 20
-------
1. Introduction
Formaldehyde emissions sampling system hardware became an issue with the
Environmental Protection Agency's (EPA) release of a Notice of Proposed Rulemaking
(NPRM) for Methanol-Fueled Vehicles and Engines1. This report discusses a
formaldehyde sampling system that EPA's Emission Control Technology Division, Test
and Evaluation Branch (TEB) will use for tests of light duty vehicles and light duty
trucks. It also discusses the considerations involved in selecting the specific hardware
used to meet TEB's needs.
The purpose of this report is to provide information on one approach for sampling
formaldehyde emissions and to explain why this approach was taken.
2. Summary
TEB found that using a thermal mass flow controller in a system to sample vehicle
formaldehyde emissions has three major advantages over a critical flow venturi. The
thermal mass flow controller:
1. allows proportional sampling with a heated or unheated probe;
2. allows the flow rate to be easily varied; and,
3. is less restrictive to flow
Additionally, the flexibility needed to allow using larger low restriction
formaldehyde collectors is enhanced by using a twin diaphragm sample pump installed
downstream of the controller.
1 "Proposed Emission Standards and Test Procedures for Methanol-Fueled Vehicles
and Engines, Draft Regulations," U.S. Environmental Protection Agency, Office of
Mobile Sources, Emission Control Technology Division, Summer 1986.
-------
3. The NPRM Required Sampling System
To comply with the NPRM, a formaldehyde sampling system must meet the
following basic requirements:
1. It must be used with a constant volume sampler (CVS) which mixes ambient air
with the vehicle exhaust and measures the total volume of the resulting dilute exhaust.
2. The formaldehyde sample flow must be taken from the dilute exhaust flow (CVS
flow) and the sample flow rate must be proportional to the CVS flow rate.
j
3. The total formaldehyde sample volume must be measured.
4. The formaldehyde sample probe and sample lines must be heated and maintained
at 235°F ±15°F. (The final regulation may not require the sample probe to be heated.)
4. Formaldehyde Sampling System Design Considerations
Section 4 presents three issues that arose in designing a formaldehyde sampling
system to be used with a critical flow venturi (CFV) type CVS while meeting the
requirements of the NPRM.
4.1 A Basic System for Formaldehyde Sampling
The system in Figure 1 illustrates a formaldehyde sampling system that can be
added to a CFV-CVS. The sample probe in Figure 1 holds a small CFV that is located
just upstream of the dilute exhaust CFV (not shown). The sample probe's CFV is used to
meet the NPRM's requirement that the sample mass flow be proportional to the dilute
exhaust mass flow. The rest of the components in Figure 1 should be self-explanatory.
-------
Heated Sample Fiobe
with Critical Flow Venturi
in Dilute Exhaust Stream
Solenoid
Valve
Heated to 235 ±15°F
Formaldehyde
Collector
Sample
Pump
Drying Tube
Figure 1
This CFV system did not meet TEB's needs for three reasons that will be discussed
in Sections 4.2 through 4.4. Section 5 will discuss the thermal mass flow controller
system that was used to overcome the problems discussed below.
4.2 The Proportional Flow Problem When Using a Heated Probe
The NPRM required the formaldehyde sample CFV to be heated to 235 ±15°F.
However, to maintain a sample flow rate that is proportional to the CVS flow rate, the
temperature and pressure at the inlets of the dilute exhaust CFV and the sample CFV
have to be the same or change by the same magnitude. Heating the sample CFV to 235°F
and controlling it there isolates the sample CFV from temperature changes at the dilute
exhaust CFV. The sample flow rate consequently does not respond to exhaust
temperature excursions as the CVS venturi does, and proportional flow is not maintained.
Testing performed after the NPRM was published confirmed that proportional flow could
not be maintained when the sample CFV was heated.^ If the final regulation drops the
requirement for a heated probe, this problem will not be relevant.
2 Adams, Walter A. "Use of Heated Crtitical Flow Venturi Sample Probes To Maintain
Proportional Flow," U.S. EPA, Report No. EPA-AA-TEB-87-01, Office of Mobile
Sources, Emission Control Technology Division, Test and Evaluation Branch, February
1987.
-------
In summary, TEB's need to comply with the NPRM requirements for proportional
sample flow through a heated probe was the first problem encountered in using a CFV for
formaldehyde sampling.
4.3 The High Pressure Drop Problem
One of TEB's design objectives was to use 2,4 dinitrophenylhydrazine (DNPH)
coated Sep-Pak silica cartridges-^ rather than impingers for formaldehyde collection.
The cartridges are easier to handle and much smaller (1 cm outside diameter by 2 cm
long) than impingers.
The pressure drop across two cartridges in series is 8.7 pounds per square inch (psi)
when flowing 1 standard liter per minute (slpm). Gabrysiak4 found that the pressure
drop across a CFV used with the Sep-Pak silica cartridges made it impractical to use a
CFV. Gabrysiak's calculations showed that using one Sep-Pak cartridge with a CFV
requires a CFV outlet pressure of 1.9 pounds per square inch absolute (psia) with the inlet
pressure at 14.7 psia. His calculation assumed that the CFV required an outlet to inlet
pressure ratio (POut/Pin) °f 0-60 or less to achieve critical flow conditions. This pressure
ratio assumption was based on an unpublished study by Carl Ryan and Bill Harbowy on
the sample CFVs used at EPA's Motor Vehicle Emission Laboratory for bag samples that
was cited by Paulina 5. The study showed that Pout/Pin had to be 0.60 or less to
maintain critical flow.
3 Tejada, Silvestre B. "Evaluation of Silica Gel Cartridges Coated In Situ with
Acidified 2,4-Dinitrophenylhydrazine for Sampling Aldehydes and Ketones in Air,"
U.S. EPA, International Journal of Environmental Analytical Chemistry, Vol. 26, pp.
167-185,1986.
4 Gabrysiak, John. "Heavy Duty Methanol Engine Test Development Project
Overview: Briefing for the Branch Chief," U.S. EPA, Standards Development and
Support Branch, January 21, 1988.
5 Paulina, Carl. "Non-Propotional Sample Rates in a Critical Flow Venturi Constant
Volume Sampler: Effects on Federal Emission Test Fuel Economy," Engineering
Operations Division, U.S. EPA, Report No. EPA-AA-EOD-84/2, January 1982.
-------
Better performing CFVs allow a POut/Pin as hi§h as ®-%® m the 1 slPm ran§e-
Figure 2 shows the calculated pressure drops across a system using two cartridges in
series with a CFV permitting a Pout/Pin of 0.80, and using Ann Arbor's average
atmospheric pressure of 14.3 psia. These conditions reduce the vacuum pump's inlet
pressure to only 2.7 psia at 1 slpm, without considering the pressure drops through the
control valves and other components. This shows that even with the more efficient of the
CFVs, the pressure drop is still prohibitively high.
20% loss
8.7 psi drop @ 1 slpm
Sample
Critical Flow
Venturi
Sep-Pak Silica Cartridges
Figure 2
Additionally, the diameter of a 1 slpm CFV for this application is only .015 inches,
which makes paniculate contamination and maintenance an issue. And, since the
maximum Pout/Pin to achieve critical flow conditions is 0.80, a safety factor is needed to
ensure that critical flow is maintained. In actual use with vehicle exhaust, a maximum
Pout/Pin °f O-7^ is needed to provide an adequate safety factor, but this increases the
restriction.
In summary, TEB's decision to use the highly restrictive Sep-Pak cartridges added a
second problem to using a CFV for metering formaldehyde sample flow.
-------
4.4 Variable Sample Flow Rate
TEB often tests prototype vehicles, catalyst and non-catalyst equipped vehicles, and
also tests over wide temperature ranges (20-95°F). These factors cause formaldehyde
emissions to vary over a wider range than is expected for certification vehicles tested
under Federal Test Procedure (FTP) conditions. The quantity of formaldehyde in the
formaldehyde collector will also vary widely if the sample flow rate is held at a constant
proportion of the dilute exhaust sample flow rate.
Liquid chromatographs (LC), used to measure the formaldehyde gathered by the
formaldehyde collector, are limited to a finite range over which measurements can be
conveniently made. Naturally, as the LC's measurement range increases, its resolution
decreases. The range in the amount of formaldehyde in the collector can be narrowed by
making the formaldehyde sample flow rate adjustable. For example, if a very low
emitting vehicle is tested, it may be desirable to increase the nominal flow rate from the
commonly used rate of 1 slpm to 1.2 slpm. This 20% increase in flow increases the
quantity of formaldehyde collected. An analogous situation exists for high emitting
vehicles. Using the variable flow rate capability allows the LC to operate over a smaller
range with greater resolution. CFVs do not allow the flow rate to be varied, except by
replacement, which is not a viable alternative since they cost at least $700 each, and
changing CFVs between tests would cause delays.
In summary, TEB's potential need for variable sample flow rates added a third
problem to using a CFV for formaldehyde sampling.
5. The TEB Formaldehyde Sample System
TEB decided to use a thermal mass flow controller system to overcome the three
problems discussed in Sections 4.2 through 4.4. First, proportional sampling can be
achieved with the mass flow controller, even with a heated probe. Second, the thermal
mass flow controller system is less restrictive, and third, the nominal flow rate can be
changed by simply dialing in a new setting. TEB's formaldehyde sampling system is
shown in Figure 3 (page 20).
The standard thermal mass flow controller system has two components, an interface
module and a mass flow controller unit. Note the terminology; the mass flow controller
unit is one component of the mass flow controller system. The system refers to both
-------
components which are electrically connected with a cable. The controller unit is a
plumbing component in the sampling system, where it performs two functions. As a
sensor it measures the flow rate, and as an actuator it mechanically controls the flow rate.
The interface module contains the control circuitry of the system and includes the flow
rate set-point controls and a digital display used for monitoring the flow rate and
displaying the set-point
Proportional sampling is achieved by connecting the CVS's nominal 0-10 volt direct
current (vdc) flow rate signal to a modified thermal mass flow controller system. The
standard controller system (without the CVS connected) uses a potentiometer on the
interface module to adjust the set-point, which is sent as a 0-5 vdc input signal to the
controller unit. The actual flow rate, through the controller unit, is represented by a 0-5
vdc output signal from the controller unit which is sent to the interface module where it is
displayed. The difference between the set-point input signal and the actual flow rate
output signal "calls" for appropriate flow control to achieve agreement between the two
signals. The CVS's 0-10 vdc signal must be substituted for the interface module's 0-5 vdc
set-point signal to keep the formaldehyde sample flow rate proportional to the CVS flow
rate.
A voltage divider circuit (see Figure 4) is used to lower the CVS signal from 0-10
vdc to make it compatible with the flow controller unit's need for a 0-5 vdc set-point
signal. The interface module's set-point potentiometer is disabled and a separate
potentiometer, which is an element of the voltage divider circuit shown in Figure 4, is
substituted for the original set-point control.
I
Potentiometer
CVS Flow Rate
I
Voltage Divider Output
to Contoller Unit
Figure 4
-------
A change in function of the mass flow controller system also occurs as a result of
using the voltage divider. The interface module's digital display normally indicated the
controller unit's flow rate, and depressing the Flow/Set-Point switch caused the set-point
to be displayed. The set-point was an absolute setting. Adding the voltage divider
deactivates the ability to display a set-point, so the display normally indicates the actual
flow rate and gives an irrevalent reading when the Flow/Set-Point switch is depressed.
Adjusting the voltage divider's potentiometer changes the actual flow rate, thereby
replacing the set-point control, but unlike the original set-point adjustment, the actual
flow rate setting is no longer an absolute setting. Remember that the signal source for
controlling the "set-point" is the CVS's 0-10 vdc flow rate signal which varies with
changes in the CVS mass flow rate. The voltage divider's potentiometer can tailor this
signal in a relative sense, but unlike the original set-point adjustment, the voltage
divider's potentiometer setting will not maintain a fixed flow rate. The sample flow rate
will now change with CVS flow rate changes. So, TEB normally uses the voltage
divider's potentiometer to initially set the nominal sample flow rate, but the flow rate will
then vary proportionally with CVS flow rate changes. These modifications were
relatively simple and only involved the cable between the interface module and the flow
controller unit; neither component was internally modified.
During a test with the new system, an increase in the dilute exhaust mass flow rate
will cause the sample mass flow rate to increase proportionally from the nominal
formaldehyde sample flow rate setting (typically 1 slpm for formaldehyde sampling).
Analogous sample mass flow rate reductions are made for the more typical case where
the dilute exhaust mass flow rate decreases with the increasing dilute exhaust
temperatures that accompany warm-up during the test. In the CFV type of CVS, the mass
flow rate is inversely proportional to the square root of the CFV's inlet absolute
temperature and directly proportional to its inlet pressure, under critical flow conditions.
The TEB system uses an electronic totalizer to measure the total sample flow.
The mass flow controller unit's 0-5 vdc output signal, which is proportional to the
formaldehyde sample flow rate, is sent to a linear frequency convener (see Figure 3) that
changes the analog signal to a frequency signal. The resulting frequency signal is sent to
a liquid crystal display totalizer that displays the total flow, and is referred to as "counts."
The counts were calibrated to units of standard volume (EPA uses standard conditions of
68°F and 29.92 inches of Hg) with a wet test gas meter. The wet test gas meter's
accuracy specification is ±1/2% of true volume, making it suitable for use as a calibration
standard.
10
-------
5.1 Thermal Mass Flow Controller System Performance
The critical requirement for the thermal mass flow controller system is for it to
maintain a sample flow rate that is proportional to the CVS flow rate. But the
measurements ultimately needed to calculate emissions in mass/distance units are the
total sample flow and the total CVS flow, rather than the respective flow rates. These
totals are read from their respective digital totalizers.
The results from a test to evaluate whether the total formaldehyde flow remains
proportional to the total CVS flow as the CVS flow rate changes are listed in Table 1.
Using counts, rather than a more direct measure such as the output voltages, allows the
performance of the entire formaldehyde sampling system to be evaluated. Included are
the frequency converter, totalizer, and thermal mass flow controller system.
Table 1
Proportionality of Sample Flow Counts to CVS Counts
Ratio of Form.
CVS Formaldehyde Formaldehyde CVS Controller
- Calibrator Settings - Total Controller Controller to Flow Flow
Pressure Temp. Time Flow Total Flow CVS Flow Rate Rate
(mmHg) (°F) (sees.) (counts) (counts) (counts) (scfm) (slpm)
760
760
500
500
650
70
70
32
130
70
900
900
240
240
240
5396
5401
941
860
1219
10472
10471
1863
1708
2377
1.941
1.939
1.980
1.986
1.950
359.7
360.1
235.3
215.0
304.8
1.0
1.0
0.7
0.6
0.9
Average Ratio 1.959
Standard Deviation 0.022
Coefficient of Variation 1.14%
Ideally, the ratio of the total formaldehyde sample flow to the total CVS flow
should remain constant. The results show that the ratio of formaldehyde counts to CVS
counts remained very constant, with a coefficient of variation of less than 2%, over a 40
11
-------
percent change in the CVS flow rate, which is a much larger change in CVS flow rate
than would occur in normal testing. These data show that the mass flow controller
precisely maintains a sample flow rate that is proportional to the CVS flow rate.
A CVS calibrator, rather than a vehicle, was used for the test. The calibrator allows
a wider range in CVS flow rates than can be easily achieved with a vehicle, and thereby
increases the stringency of the test. The CVS's CFV inlet pressure and temperature
sensor connectors were removed from their sensors and connected to the calibrator. The
calibrator sends simulated pressure and temperature signals to the CVS computer, so the
CVS counts do not represent the true total flow through the CVS. This is a valid method
since the computer in the CVS uses the simulated sensor signals just as it would use the
normal signals to calculate flow through the CFV. And also as in normal operation,
sends the result to the CVS totalizer and to the mass flow controller, as previously
discussed. The calibrator settings for each test point are listed in columns 1 and 2 of
Table 1.
The mass flow controller's response time to changes in CVS flow rate was also
checked with a qualitative test One channel of a chart recorder was connected to the
CVS's 0-10 vdc flow signal output and a second channel was connected to the sample
flow controller's flow signal output. The recorder was started with a vehicle at zero
speed, then the vehicle was accelerated to 50 mph. As the CVS flow decreased with
increasing dilute exhaust temperature, the sample flow tracked the CVS flow very
precisely. This was true even for some rapid flow rate transients of small amplitude.
In summary, the mass flow controller system was checked to determine if it can
maintain proportional sample flow as the CVS flow rate changes and the data indicated
that it does.
5.2 Flow Controller and Sample Pump Selection Interactions
This section discusses three issues regarding flow controllers and sample pumps.
They include: 1) the operating pressure of the flow controller, 2) whether the pump
should be located upstream or downstream of the flow controller, and 3) the pump
selection.
The mass flow controller unit is calibrated at the factory for the intended operating
pressure, so the operating pressure should be specified when ordering the controller.
Determining the operating pressure was difficult because the pressure is determined by
12
-------
both the sample pump's characteristics and the pressure drop across the formaldehyde
sampling system's components.
Simply selecting the pump before specifying the controller pressure appears to be
the solution, but as discussed below, other factors had to be considered. The sample
pump and the flow controller interact. As the mass flow controller operating pressure is
dependent on the sample pump, so too is the pump selection dependent upon the pressure
drops in the mass flow controller and the other sample system components.
An orifice in the flow controller unit causes a significant pressure drop, thus
reducing the sample flow rate for a given pump. Uince the controller manufacturer uses
different sized orifices, depending on the specified operating conditions. This interaction
between the pump and the controller must be considered when selecting a pump and
specifying the calibration pressure of the controller.
Another issue is whether the pump should be located upstream or downstream of
the controller. Some manufacturer's controllers will not operate under vacuum conditions
(pump downstream). Porter Instrument Company recommended using their controller
with an upstream pump, even though the controller can operate under vacuum conditions.
One reason is that the controller is calibrated at a specific pressure, hence a measurement
error will occur at any other operating pressure.
Measurement errors can be avoided by placing the pump upstream of the controller
with a pressure regulator installed at the controller's inlet. The pressure regulator
maintains a constant pressure at the controller's inlet despite varying flow rates, thus
meeting the need to maintain the calibration pressure.
Measurement errors can not be avoided with the pump downstream of the
controller. As the controller changes the pressure drop across itself to regulate the flow
rate, the pressure at the inlet of the controller also changes. Nothing analogous to a
pressure regulator was found that can keep the pressure at the inlet of the controller
constant while allowing the flow rate to change. So a measurement error will occur at
any pressure other than the calibration pressure, if the pump is downstream of the
controller.
Initially the controller manufacturer's recommendations were followed with the
pump located upstream of an on-hand controller. With an upstream pump, leaks can not
be tolerated because the flow controller system is used to measure the total flow through
13
-------
the cartridges. Carbon vane pumps and piston pumps typically have small leaks and
metal bellows type pumps are expensive, leaving diaphragm pumps as the best
alternative. Using an on-hand single diaphragm pump, the maximum flow rate was
limited io 0.75 slpm. As stated earlier, our objective was to attain a nominal sampling
rate of at least 1 slpm with the capability to use higher sampling rates if necessary. The
upper limit was 5 slpm since thermal mass flow controllers are commonly available as 0-
5 or 0-10 slpm units, and there was not a need for anything more than 5 slpm..
In looking for a suitable pump, the pump manufacturers recommended locating the
pump downstream of the flow controller, directly contrary to the controller
manufacturers' preference for placing the pump upstream of the controller. The decision
became a matter of choosing the configuration which had the fewest drawbacks.
The following discusses the problems that arise with the pump located upstream of
the controller. Diaphragm vacuum pumps are not designed to operate with pressures
above atmospheric at their outlets. If an oversized pump is used, the mass flow controller
is forced to restrict the pump outlet, thus creating a positive pressure at the outlet. Too
much restriction will overload the pump and motor, causing damage, while an undersized
pump will fail to deliver the required flow rate. In effect, any pump will be oversized
when prevented from delivering full flow by restricting the outlet. These considerations
severely restrict the range in flow rates than can be achieved without causing damage.
A second pump durability problem arises with the pump located upstream of the
controller. An upstream pump must simultaneously pull a vacuum across the cartridges
and deliver a pressure at the inlet of the controller to overcome the pressure drop caused
by the orifice and the flow regulating valve in the controller. But diaphragm pumps are
designed as either vacuum or pressure pumps exclusively. The pump manufacturers said
that pulling a vacuum at the inlet of the pump while simultaneously pushing a positive
pressure at the pump's outlet adversely affects the diaphragm's durability, so is not
recommended. The problem occurs during the transitions between positive pressure and
negative pressure acting on the top surface of the diaphragm while atmospheric pressure
is constantly acting on the diaphragm's underside. During the transitions, the diaphragm
is snapped up when exposed to vacuum and snapped down when exposed to pressure.
For these reasons, the pump manufacturers do not publish flow rate specifications
for conditions where the pump is simultaneously exposed to less than atmospheric
pressure at the inlet and greater than atmospheric pressure at the outlet. Choosing a pump
for an upstream location is practically impossible.
14
-------
In addition to the durability issues, another disadvantage of installing the pump
upstream of the controller is that the pump is the component most likely to develop a
leak. A leak in an upstream pump would cause inaccuracies in the total flow volume
measurement. A leak in a downstream pump would not cause measurement inaccuracies.
Another advantage of a downstream pump is that it enhances the flexibility of the
sampling system. As discussed in Section 4.3, the Sep-Pak silica sampling cartridges are
very restrictive to flow. TEB is looking into less restrictive cartridges and if successful,
they will require the controller unit to further restrict the sample flow to maintain a
constant flow rate. With the controller on the inlet side, the pump would not "see" a
difference in restriction since the sum of the cartridge restriction plus the controller
restriction will not change while maintaining a constant flow rate.
With the controller on the outlet side of the pump, the outlet pressure will increase
in proportion to the lowered cartridge restriction. The increased pump loading resulting
from a downstream controller restricting the pump outlet could cause the motor and/or
pump to overheat. So less restrictive cartridges might require changing to a smaller pump
on systems with upstream pumps. A downstream pump does away with these problems
and makes the system more accommodating of component changes such as less
restrictive cartridges.
Although there are significant advantages associated with a downstream pump,
there are also two disadvantages. First, some manufacturers' controllers will not work
under vacuum conditions thus reducing the number of suppliers, and second, flow rate
measurement errors accompany operation at pressures other than the controller unit's
calibration pressure.
The controller unit's calibration pressure turned out to be a minor issue. The Porter
Model 201 0-5 slpm controller unit's pressure coefficient specification is
0.1%/atmosphere, so the maximum error would only be 0.1% if the controller is
calibrated at atmospheric pressure even if a perfect vacuum were attained at the inlet.
The error is even less if the controller is calibrated near the middle of its expected
operating pressure range. It is also a simple matter to calculate a corrected flow rate
although it requires measuring the difference between the calibration pressure and the
operating pressure. The anticipated error is so insignificant that it would be difficult to
justify even such a minor effort.
15
-------
5.3 Flow Controller and Sample Pump Selection Decisions
The advantages of placing the pump downstream of the controller outweighed the
disadvantages. A downstream pump lowers the probability of leaks, favorably affects
system flexibility, and enhances pump durability.
A Cast 1/4 hp, 1575 rpm twin diaphragm pump model number DAA-155-EB was
selected. This pump, and similar pumps manufactured by others, are available in series
and parallel diaphragm configurations and they are easily changed from one
configuration to the other, in the field. This can be an important consideration if lower
restriction cartridges become available. At the high vacuums required with the present
cartridges the series diaphragm configuration allows higher flow rates. If lower
restriction cartridges become available, they may also be larger. In order to maintain
equivalent sample concentrations with larger cartridges, higher flow rates will be
necessary. The parallel diaphragm configuration delivers higher flow rates than the series
diaphragm when the system restrictions are lessened.
Table 2 lists the pressures at the inlet of the controller using the Gast pump with a
mass flow controller and two Sep-Pak silica cartridges in series.
Table 2
Flow Controller Inlet Pressures at Various Flow Rates
With Two Sep-Pak Cartridges in Series Upstream of Controller
Controller
Flow Inlet Barometric
Rate Pressure Pressure
(slpm) (psia) (psia)
1.21 2.01 14.29
1.00 5.55 14.29
0.88 7.20 14.29
0.75 8.57 14.29
0.70 9.09 14.29
16
-------
5.4 Additional Details of the TEB System
This section provides information on the other components in Figure 3 that'
complete the sampling system.
The first component considered are the three way heated solenoid valves just
downstream of the sample probe. TEB's formaldehyde sampling system will share the
sample probe with a methanol emission sampling system. The design of the methanol
sampling system has not been finalized and is therefore not covered in this document.
However, some of the components depicted in Figure 3 were selected to accommodate
the methanol sampling system. The three way solenoid valves, for instance, would have
been a two-way valve were it not for the methanol system.
The water used in the mode 1 methanol impingers can be sucked out when the test
driver switches from mode 1 to mode 2 of the FTP. The high vacuum on the
formaldehyde side of the combined formaldehyde/methanol system creates a vacuum
reservoir that pulls water out of the impingers just after the mode change. With three-
way solenoid valves, the downstream port opens to atmosphere when sampling stops at
the mode change and prevents the formaldehyde circuit from pulling water out of the
impingers.
The two valves just downstream of each of the heated solenoid valves allow the
methanol sampling branch to be closed when only formaldehyde sampling is desired or
vice versa. These are manually set by the technician prior to the test.
The three downstream mode valves allow one pump and controller to be used for
the three modes of the FTP. Only one of these valves is open during each mode of the
test, thus isolating the rest of the system from the pump.
The ambient solenoid valve and the needle valve were added to aid the accuracy of
the total flow volume measurement and enhance pump durability. The following
operating sequence explains why. The pump and flow controller system are turned on
before a test begins to set the nominal flow rate and to allow a ten minute warm-up, as
recommended by the controller manufacturer. The ambient solenoid valve is normally
open and allows air to pass through the controller and pump. The downstream solenoid
valves are normally closed and remain closed until the driver selects one of the test
modes. The ambient solenoid valve thus prevents the pump from dead-heading which
aids durability and permits air flow through the controller for setting the flow rate.
17
-------
The needle valve improves the accuracy of the total flow measurement. When the
driver selects mode 1, simultaneously the ambient solenoid valve closes, the mode 1
downstream solenoid valve opens, the mode 1 heated solenoid valve is energized to open
the sample probe to the cartridges and close the ambient port, and the totalizer is
energized and begins counting.
The sampling system pressure now begins decreasing from atmospheric pressure.
The lines between the cartridges and the pump are evacuated to approximately 6 psia.
Attaining this low pressure causes more air to flow through the controller unit than flows
through the cartridges. Since only the flow through the cartridges is pertinent, this
additional flow is being counted and becomes an error source. The counter is only
actuated when one of the downstream solenoid valves is energized.
This error is minimized with the needle valve. By adjusting the needle valve to
create the same pressure drop as the cartridges, the entire sampling system between the
pump and the downstream solenoid valves can be evacuated to the sampling pressure, so
the counters measure the flow through the cartridges without the additional flow required
to evacuate the system. The distance between the cartridges and the downstream solenoid
valves must be minimized to reduce the error associated with evacuating this volume.
The filter shown in Figure 3 is needed to protect the capillary tube in the flow
controller unit from condensed water and solids. A problem with condensed water is
unlikely under normal operating conditions since the pressure is so low at the inlet of the
controller, but solids are still a potential problem. For example, a leak in the downstream
solenoid valves of a prototype system was caused by silica from the cartridges. The
silica, held by fine mesh filters, can escape unless care is taken to prevent breaking the
seal between the cartridge's body and filters during installation. Also, during non-routine
procedures such as repairs, the operating conditions could be amenable to water
condensation or solids contamination, which further justifies the filter.
18
-------
6. List of Suppliers
Thermal Mass Flow Controller
Model 201 0-5 slpm
Sample Pump
Model DAA-155-EB
Solenoid Valves
Totalizer
Model 44611
Frequency Convener
Model ST-30
Filter
Model A912-DX with
manual drain valve
Sep-Pak Silica Cartridges
Porter Instrument Co.
P.O. Box 326
Hatfield, PA 19440
215-723-4000
Cast Manufacturing Corp.
P.O. Box 97
Benton Harbor, MI 49022
616-926-6171
Peter Paul Electronics Co.
P.O. Box 1180
New Britain, CT 06050-1180
203-229-4884
Eaton Corporation
901 South 12th Street
Watertown.WI 53094
414-261-4070
RC Systems
P.O. Box 57338
Webster, TX 77598
409-925-7808
Balston, Inc
P.O. Box C
Lexington, MA 02173
800-343-4048
Waters Chromatography Division
Millipore Corporation
34 Maple Street
Milford,MA 01757
19
-------
Figure 3: Formaldehyde Sampling Syslem
Sample Probe in
Dilute Exhaust Stream
in
iream
S
Heated to
235 ±15°F
Nk
i-*^D*b-I
%.
Mode 3
Heated
Solenoid
Valve
NcJ-7 f~
~^"^1
Mode 2
Heated
Solenoid
Valve
CMO
Model ^
Heated
olenoid Valve
#76PP9ZGV
Mode 3
To methanol Downstream
~ impingers Solenoid
Valve
O O I^1^
Sep PAK
Silica
Cartridges
Mode 2
To methanol Downstream
~ impingers Solenoid
Valve
Sep PAK
Silica
Cartridges
Model
To methanol Downstream
impingers Solenoid Valve
#22N7DGV
s~~~\ /*~~*\ r^_^^
Sep PAK
Silica
Cartridges
( Analog Flow j
Signal 1
from CVS 1
I Eaton-E
Ponerlnierface / RC Systems ModeN
Module j Model ST-30 Totaliz
"\ f X Frequency GasVc
\lf Converter Read
Balston#A912-DX |
Coalescing hiltcr
^h r 1 ^\ r^^^^^^"^~^^l ^k f
^^ V 1 f \ \ "^S- ~f
^^r -• ^
Porter Model 201 Cast Model
0-5 slpm Mass DAA-155-EB
Flow Controller Unit Vacuum Pump
Needle
. Valve
Ambient
Solenoid
Valve
#71P9ZGM
Pidgeon: 3/25/88
Revised: 6/21/88
------- |