Does duct smoke detection really work? Is it worth the cost? Is the potential for false and nuisance alarms worth the added protection? The universal answer to all fire protection questions is: It depends. This is the second part of a two-part article that summarizes the findings of research recently conducted under the auspices of the Fire Detection Institute. The first part, published in the Winter 2006 issue, summarized the purpose of the research and findings in the areas of smoke-driving forces, smoke dilution in ductwork, and the effects of smoke aging on detection in ductwork. This part summarizes findings regarding the effects of HVAC filters on smoke detection in ducts, smoke stratification in ducts, and the efficacy of duct detectors that use sampling tubes.

Effects of HVAC Filters
Smoke consists of solid particles, liquid droplets, gases, and agglomerates of these three classes of matter. As expected, filters in a mechanical ventilating system remove a portion of the smoke as the contaminated air passes through them. This effect varies depending upon the type of filter that is in place in the HVAC system.


The performance of filters for mechanical ventilation systems is quantified through testing in accordance with ANSI/ASHRAE 52.11 or ANSI/ASHRAE 52.22 The measure that is most appropriate for smoke is called "dust-spot efficiency." The ASHRAE standards group filters on the basis of the range of dust-spot efficiency they achieve in standard tests. Table 1 summarizes the grouping by dust spot efficiency and construction, and Figure 1 compares filter capabilities to typical particulates.


Studies of smoke particles have shown that they usually have a size within the range of 0.4 to 3.0 microns, depending upon the fuel and the mode of combustion (flaming versus smoldering).3


Testing in the full-scale NRC facility showed that the HVAC filters did exert an important effect on duct smoke detector performance. Group 1 filters with a dust-spot efficiency of 10 percent to 15 percent constructed of glass fibers in a cardboard frame reduced the analog output of ionization detectors an average of 20 percent, while it reduced the analog output of the photoelectric detectors by 35 percent. Group 2 filters (constructed of extended area, pleated wet-laid cellulose fiber) with a dust spot efficiency of 30 percent to 35 percent reduced the analog output of ionization detectors by about 40 percent while it reduced the analog output of the photoelectric detectors by about 55 percent.



These tests confirmed the effects of filters on smoke detector response. As expected, the effect is greater on photoelectric smoke detectors than on ionization detectors, which tend to be more sensitive to small-diameter particles that pass through the filters. However, examination of Figure 1 shows that more efficient filters can also remove smaller-sized particles and would be expected to affect the response of ionization sensors as well as photoelectric units.


Filter position, type, and efficiency must be considered when evaluating the expected performance of duct smoke detectors. (See Figure 2.) In most air return ducts, the duct smoke detector is placed after all air inlets, and before any filters, fresh air inlets or fans. Therefore, filters would not be a factor. On the supply side, the duct smoke detector is placed after the fresh air inlet, filter, conditioning area, and fan. Thus, filters would be a factor for detecting fires other than a filter fire itself.



Stratification in Ducts
NFPA 72 recommends (but does not require) that duct smoke detectors be located in a duct section that is between 6 and 10 equivalent duct diameters from bends or openings. (The equivalent duct diameter for a rectangular duct is the diameter of a circle that results in the same area as the rectangle.) It also suggests that long uninterrupted straight runs of duct may cause stratification within the duct and thus affect duct detector placement or response.


The UMDFPE team used the medium-scale test duct and cone calorimeter to investigate the possibility of stratification or nonuniform smoke distribution at one point in a horizontal duct. The data showed that at low velocities, buoyancy causes the smoke to concentrate in the upper portion of the duct. As the velocity increases, the distribution becomes more uniform. The degree of stratification depends on the temperature of the smoke relative to the ambient air and the velocity in the duct.



The NRC team installed duct smoke detectors on the side of a long section of duct along the duct centerline with horizontal sampling tubes and also on the bottom of the duct, along the duct centerline, with vertical sampling tubes. (See Figure 3.)


The NRC test results show very similar performance for all of the detectors on the duct, regardless of their location or orientation. It should be noted that the test location in the NRC test program was 10 stories above the fire test room. Consequently, the smoke had had ample opportunity to cool as it traveled from the fire test room (first floor) to the HVAC mechanical room (tenth floor), and there was no observable thermal stratification as observed in the work performed by UMDFPE where the data were collected in a duct section close to the fire source.



These NRC results suggest that there is no justification for the 6 to 10 duct diameter rule with respect to turns in the duct when the detector is located a long distance from the fire source. However, the distance necessary to result in even mixing and no stratification is not defined. The UMDFPE test results clearly show that smoke with any appreciable buoyancy results in upper-level stratification of the smoke. Thus, for any given smoke temperature above ambient, it is expected that there will be stratification in the duct until it travels some distance and mixes with ambient air and loses energy to the surrounding environment. The distance where stratification will be a factor is dependent on the temperature of the smoke relative to ambient air.

Combining the NRC and UMDFPE results suggests that mounting duct smoke detection in the upper part of horizontal ducts is preferred. Duct smoke detection utilizing sampling tubes may also be installed vertically. However, with large ducts, vertical mounting may result in slower response-as lower parts of the sampling tube will be exposed to cleaner air, diluting the sampling from the stratified layer in the duct.


Efficacy of Duct Detectors Using Sampling Tubes
There are two types of duct smoke detection. The first uses spot-type smoke detectors installed in the ducts the same types of detectors that are installed on the ceilings in a building. The second type is a similar detector in a housing with sampling tubes that enter the duct to create a flow of sampled air through the detector. Both must be tested at a range of velocities to ensure that they operate properly.


There has been concern expressed in the past that current mechanical ventilating systems operate at flow velocities that are outside the range for which duct smoke detectors are tested and listed. The testing of duct smoke detectors by Underwriters' Laboratories, Inc., includes five flow velocities: 1.52 m/sec (300 ft./min), 5.08 m/sec (1000 ft./min), 10.16 m/sec (2000 ft./min), 15.24 m/sec (3000 ft./min), and 20.32 m/sec (4000 ft./min).4


The UMDFPE team surveyed a total of 65 buildings in the Baltimore/ Washington, D.C., area, recording the installation details regarding the duct smoke detectors. The measured velocities ranging from 2.06 m/sec (405 ft./min) to 40.64 m/sec (8000 ft./min). However, all but two of the facilities had systems with velocities below 20.32 m/sec (4000 ft./min).


Thus, all but two of the 65 surveys had velocities within the current test range.


The NRC team measured the response of duct smoke detectors as a function of system air velocity. The range of velocities NRC tested went from a low of 4.0 m/ sec ( 787 ft./min) to 19 m/sec (3740 ft./min). Over this range, no significant variation in detector performance was observed.



  1. ANSI/ASHRAE 52.1, Gravimetric and Dust-Spot Procedures for Testing Air-Cleaning Devices Used in General Ventilation for Removing Particulate Matter, American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, 1992.
  2. ANSI/ASHRAE 52.2, Method of Testing General Ventilation Air-Cleaning Devices for Removal Efficiency by Particle Size, American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, 1992.
  3. Mulaolland, G., "Smoke Production and Properties," SFPE Handbook of Fire Protection Engineering, 3rd Edition, National Fire Protection Association, 2002, Chapter 2-13.
  4. UL 268A, Smoke Detectors for Duct Applications, Underwriter's Laboratories, Inc., Northbrook, IL, 1995.

Editor's Note About This Article
This is a continuing series of articles that is supported by the National Electrical Manufacturer's Association (NEMA), Signaling Protection and Communications Section, and is intended to provide fire alarm industry-related information to members of the fire protection engineering profession.