has not been reviewed or approved by, and may not necessarily reflect the views of,
the Commission.
2
COMBUSTION SENSOR TEST RESULTS



I.
EXECUTIVE SUMMARY ................3
II. INTRODUCTION
..................5
Hazard .....................5
Background .................5
Objectives ...................7
III.
SENSOR DESCRIPTION ..................................8
Electrochemical Carbon Monoxide Sensor ................8
Infrared Carbon Dioxide Sensor .................8
IV.
CHARACTERIZATION TEST RESULTS AND DISCUSSION .....................9
Test Chamber and Data Acquisition ...................9
Response Test Discussion .........................10
Electrochemical CO Sensor Response Tests ................................................11
Infrared CO
2
Sensor Response Tests ................17
Interference Tests ..................21
Electrochemical CO Sensor Interference Test Results .................22
Infrared CO
2
Sensor Interference Test Results .................................................25
V. FURNACE
SENSOR
SHUTOFF TESTS AND DISCUSSION .....................26
Test Chamber and Flue Gas Sampling ..........26
Shutoff Circuit to Furnace ........27
Electrochemical Carbon Monoxide Sensor Shutoff Test Results ............29
Flue Gas Profile Tests ...................35
Infrared Carbon Dioxide Sensor Shutoff Tests .....37
VI. SUMMARY
......41
VII. CONCLUSION
.....42
ACKNOWLEDGEMENTS ..........43
REFERENCES .....44
APPENDIX A. TEST MATRICES ......45
APPENDIX B. ADDITIONAL DATA ....53
ENDNOTES .....56



3
I. EXECUTIVE
SUMMARY
This report describes work performed by the U.S. Consumer Product Safety Commission (CPSC)
staff in 2003 to evaluate sensor technologies to demonstrate their ability to provide detection and
shutdown response to carbon monoxide (CO) concentrations in excess of 400 parts per million
(ppm) within the vent system or flue passageways of a residential gas furnace. This test program
was a primary task within the Vented Gas Appliance CO Sensor Project and was an extension of
sensor evaluations conducted by CPSC staff in 2001.

One of CPSCs three strategic goals is to reduce the rate of death from CO poisonings associated
with consumer products by 20 percent from the 1999-2000 average by the year 2013 (1999-2000
average yearly estimated CO poisoning deaths from unintentional non-fire consumer product-
related incidents was 124). The goal of this test program was to support development of a
performance standard that would require shutdown or some other preemptive response to elevated
levels of CO within the flue passageways of vented gas heating appliances (e.g., residential
furnaces, boilers, and room heaters). The intended outcome is a reduction in CO-related deaths
caused by gas heating appliances.

Gas heating systems are the leading cause of CO poisoning deaths associated with consumer
products. From 1999-2001, there was an average yearly estimated 126 unintentional non-fire CO
poisoning deaths associated with consumer products. Heating systems of all types were associated
with 69 deaths, or 65 percent of the CO poisoning deaths associated with consumer products
excluding engine-powered tools. Among heating systems, from 1999- 2001, natural gas heating
was associated with an average yearly estimate of 28 deaths (41% of the heating system deaths); LP
gas heating was associated with an average yearly estimate of 26 deaths (38% of the heating system
deaths); and unspecified gas heating was associated with an average yearly estimate of 5 deaths (7%
of the heating system deaths). Overall, an average yearly estimate of 59 CO poisoning deaths from
1999-2001 were associated with the use of gas heating systems.

The objectives of this test program were to: evaluate gas sensor performance under various
humidity and temperature conditions; evaluate indirect measurement and response to CO using a
carbon dioxide (CO
2
) sensor; evaluate sensor performance in the presence of potential interferent
gases; and evaluate shutdown response from each sensor while integrated into a furnace. This
sensor evaluation was comprised of characterization testing within an environmental chamber and
in-situ furnace shutoff testing. The characterization testing examined each sensors performance
when exposed to target gases and non-target gases under a variety of conditions. Sensor
performance observed during the characterization testing was used as a basis for selecting target
parameters for in-situ furnace shutoff testing. For in-situ shutoff testing, each sensor was separately
integrated into the test furnace and subjected to testing of varying durations.

Two different gas sensor technologies were evaluated. One technology used an electrochemical
sensor with a target gas response range of 0 to 1000 parts per million (ppm) CO. The other
technology used an infrared sensor with a target gas response range of 0 to 20.6% CO
2
. The
infrared carbon dioxide sensor was selected for testing on the basis that changes in flue
concentrations of CO
2
might be used as an indicator of an increase in the flue concentration of CO
in excess of the 400 ppm (air free) emissions standard for residential gas furnaces.

4
All of the objectives of this test program were accomplished. Each sensor exhibited a direct, linear
response to its respective target gas and did not respond to non-target gases under various test
conditions. The test results also demonstrate each sensors capability to shutdown the furnace in
response to elevated concentrations of CO within the flue. These results are limited to conditions
exhibited by a high-efficiency gas furnace.

Issues such as sensor reliability, durability, expected life, and performance in higher temperature
environments (e.g., 300
o
F to 500
o
F) were not addressed by this test program. Future test and
evaluation of sensors should encompass a wider variety of sensor technologies, target gases, and
exposure to potential contaminants. These issues are addressed in a draft test matrix developed by
CPSC staff for the Canadian Standards Association (CSA)/American National Standards Institute
(ANSI) Z21/83 Ad Hoc Working Group for CO/Combustion Sensors. The test matrix is part of a
work plan to evaluate sensor usage in gas appliances developed by this working group for
consideration by the CSA/ANSI Z21/83 Technical Committee. The CPSC staff will provide this
test report to the technical committee, ad hoc working group, and the CSA/ANSI Z21.47 Central
Furnace Technical Advisory Group to further support sensor evaluation and development of a
performance standard.
5
II. INTRODUCTION

Hazard
Carbon monoxide is a by-product of the incomplete combustion of hydrocarbon fuels such as
natural gas, propane, gasoline, and oil. Incomplete combustion from gas-fired appliances, such as
furnaces, boilers, and wall heaters, can occur as a result of an improper fuel-air mixture to the
appliance burner, quenching of the burner flame, or over-firing of the appliance above its design
energy input rate. An improper fuel-air mixture can occur as a result of a reduction or stagnation of
the primary and secondary air supplied to the burner (such as might occur when an appliance vent
pipe is partially blocked or when the appliance is installed in an undersized room). An improper
fuel-air mixture can also occur as a result of an excessive gas manifold pressure. When the flue
passageways and venting systems of appliances are intact, CO that results from incomplete
combustion is safely vented to the outdoors. However, CO can enter the living space and create a
hazard to consumers when a leakage path is created by a compromised flue passageway or venting
system.

One of CPSCs three strategic goals is to reduce the rate of death from CO poisonings associated
with consumer products by 20 percent from the 1999-2000 average by the year 2013 (1999-2000
average yearly estimated CO poisoning deaths from unintentional non-fire consumer product-
related incidents was 124).
1
The goal of this test program was to support development of a
performance standard that would require shutdown or some other preemptive response to elevated
levels of CO within the flue passageways of vented gas heating appliances (e.g., residential
furnaces, boilers, and room heaters). The intended outcome is a reduction in CO-related deaths
caused by gas heating appliances.

Gas heating systems are the leading cause of CO poisoning deaths associated with consumer
products. From 1999-2001, there was an average yearly estimated 126 unintentional non-fire CO
poisoning deaths associated with consumer products.
2
Heating systems of all types were associated
with 69 deaths, or 65 percent of the CO poisoning deaths associated with consumer products
excluding engine-powered tools.
3
Among heating systems, from 1999- 2001, natural gas heating
was associated with an average yearly estimate of 28 deaths (41% of the heating system deaths); LP
gas heating was associated with an average yearly estimate of 26 deaths (38% of the heating system
deaths); and unspecified gas heating was associated with an average yearly estimate of 5 deaths (7%
of the heating system deaths).
4
Overall, an average yearly estimate of 59 CO poisoning deaths from
1999-2001 were associated with the use of gas heating systems.
5


Background
In 1996, CPSC