Back to Main Page | News | Features | Gallery | Columns | Hands-On | Backscatter | Resources

16 July 2004

Where there's smoke, there's (not always) fire

An inside look at Smoke Detectors

Mike Dziekan, Connecticut Analytical Corporation

Part 2. Photoelectric smoke detectors

Photoelectric smoke detectors can detect smoke from smoldering fires and blazes like this one. Photograph by Forrest M. Mims III.

In Part 1 (9 July 2004) we discussed flame and heat detectors. Here we discuss detectors designed specifically to detect smoke.

Photoelectric smoke detectors

The photoelectric type of smoke detector utilizes light as a detection mechanism. There are two basic types of photoelectric smoke detectors:

•  Light sensing (scattering).

•  Light obscuring (blocking).

The light sensing type relies on the scattering of light by small particles of smoke. The light obscuration type relies on the blocking of the light beam between a light emitter and a sensor.

The light obscuration (extinction) type detector (such as the projected beam type) requires that smoke particles obscure or block a portion of a light beam transmitted to a sensor. Smoke between the light source and light sensor will diminish or reduce the amount of incident light on the sensor. If the light is diminished to a low enough level, an alarm is sounded.

Light scattering smoke detectors

Light sensing photoelectric smoke detectors depend on the ability of small airborne particles to scatter light. Scattering is a complex interaction of particles with electromagnetic radiation that involves several factors, including absorption, reflection, refraction, polarization, and diffraction. Below is a simplified view of what happens inside a photoelectric type smoke detector when smoke particles enter the chamber.

Detection of smoke by scattered light. No smoke is present in the left panel. The smoke in the right panel scatters some of the light from the source toward the photosensor.

If you have used a flashlight on a very clear night, you know that the light beam is invisible when viewed from the side. However, if fog, haze or smoke is present, the beam can be easily viewed from the side. Water vapor and steam will also scatter and reflect the beam of light. This scattering is how a scattering-type smoke detector works.

When the system's chamber is clear (free from any smoke particles), the light emitted from the light source is not detected by the photosensor. When smoke particles enter the chamber, some of the light is scattered toward the photosensor. As the amount of smoke that enters the chamber increases, more incident light is scattered toward the photosensor. When the concentration of particles reaches a predetermined threshold, an alarm will sound.

Several factors influence the amount of light that is scattered. They include the wavelength and intensity of the light and the scattering angle. Also important are the size of the smoke particles and their refractive index, shape, reflectance and concentration. More information about typical kinds of background particulate matter can be found in an article entitled Fine Particle Pollution .

There are two types of scattering, Rayleigh and Mie. Rayleigh scattering occurs when the particle size is less than that of the wavelength of the light source. Because Rayleigh scattering is inversely related to the wavelength of light, blue light is scattered much more than red light. This is why Rayleigh scattering from molecules of air makes the sky appear blue.

Mie scattering is much more complex, for there are no limitations on the particle size. In smoke from a typical fire, Mie scaterring from large smoke particles provides much more scattering than Rayleigh scattering from fine particles. Thus, photoelectric smoke detectors rely much more on Mie scattering than Rayleigh scattering.

Most photoelectric smoke detectors use a near-infrared (IR) light-emitting diode (LED) as a light source. The most commonly used LEDs in this application emit invisible near-IR light having a wavelength of 880 nm or 940 nm. Silicon photodiodes respond very well to these wavelengths and often serve as photosensors in photoelectric smoke detectors.

Various kinds of lasers can also be used as light sources in photoelectric smoke detectors. One example is the System Sensor Pinnacle laser detector . Laser smoke detectors have a higher sensitivity and signal-to-noise ratio than the LED type. Before the use of special signal processing algorithms, laser smoke detectors were usually used only in very clean environments, such as a clean room in a semiconductor manufacturing plant.

In commercial fire systems utilizing intelligent fire panels, the analog signal from the photosensor is converted to a discrete value and sent to the fire panel. The fire panel has different algorithms that will interpret this data and determine if the condition warrants a high chamber value indication, a warning indication, a pre-alarm, or a full-blown alarm condition. Standalone residential smoke detectors, such as those purchased from Home Depot, or Wal-Mart, provide only an alarm and no external access to this information.

Light obscuring smoke detectors

A projected beam photoelectric smoke detector can use a single light source and photosensor to monitor a very large area. Most such detectors use a near-IR LED or laser diode as a light source.

The projected beam detector arrangement shown below uses a separate emitter and detector. There are also combination detector/emitters that use a single passive reflector. The distance "L" in the diagram below can be up to 100 meters (several hundred feet) in one commercial system..

Detection of smoke by attenuated light. No smoke is present in the left panel. The smoke in the right panel blocks some of the light arriving at the photosensor.

Summing up photoelectric smoke detectors

Photoelectric smoke detectors may respond more quickly than ionization detectors to low energy fires, such as slow, smoldering fires. The low energy fire will generally produce large smoke particles that are easily detected by optical means.

Commercial smoke detectors determine detector sensitivity as an obscuration level per unit length. A typical smoke detector might have an early alert warning level set at 0.2%/ft and a full-blown alarm level at 3%/ft. Note that these are industry standards in the United States. Elsewhere, sensitivity levels are expressed in terms of meters (1 foot is about 30 cm).

It is important to note that even if a room is filled with smoke having a 2.5%/ft obscuration level, a detector set to trigger at this threshold may not immediately respond. The smoke must first enter the internal sensing chamber or the path of the light beam. A detailed study was done on particulate entry lag time in spot type smoke detectors .

Another class of photoelectric smoke detectors are known as aspirated smoke detectors . Aspirated sensing systems use a piping manifold to connect sensing tubes to a central air pump. The pump draws air, and any smoke it contains, into the detector sampling chamber. This type of system is great for locations where it is convenient to run lengths of small sensing tubes perforated with tiny sensing holes. This system also has the advantage of concentrating an air sample that may be diluted with clean air if sensed by a spot type detector.

Another similar type of detector is the "duct detector." This type of smoke detector monitors the HVAC air traveling through building ductwork. It has the advantage of using a single detector to sample a much greater volume of air than would be possible with an ordinary spot type smoke detector. Aspirated and duct detector systems utilize a single spot type photoelectric detector. Why an ionization type detector is not used will be addressed in Part 3 of this series.

It is important to know that photoelectric smoke detectors will also respond to particles besides smoke. I have seen more than one detector accidentally triggered by a high pollen count or by someone using an old vacuum cleaner to pick up some dust or powder.

For those wishing to delve deeper into the science of aerosol measurement, I recommend "Aerosol Measurement -- Principles, Techniques, and Applications," which is edited by Klaus Willeke and Paul Baron (ISBN 0-442-00486-9). Because this book costs about $200 new, I suggest buying a used copy. If you can afford this book, I assure you that it will be a frequently accessed source when dealing with aerosol measurement techniques.

Many papers discuss the structure and scattering properties of smoke particles emitted by incipient and fully flaming fires. Some interesting papers can be obtained from NIST (National Institute of Standards and Technology). Several key papers are listed below:

International Conference on Automatic Fire Detection "AUBE '01", 12th. Proceedings . National Institute of Standards and Technology. March 25-28, 2001. Gaithersburg, MD.

Apparatus for Light Scattering Studies of Smoke Particles by Weinert, D. W.; Mulholland, G. W. NIST SP 965; February 2001.

Size Distribution and Light Scattering Properties of Test Smokes by Weinert, D. W.; Cleary, T. G.; Mulholland, G. W.

Light Scattering Characteristics and Size Distribution of Smoke and Nuisance Aerosols by Weinert, D.W., Cleary, T. G., Mulholland, G. W., Beever, P F.

Here are some articles by amateur scientists on the atmospheric scattering of light:

Shawn Carlson, "When Hazy Skies Are Rising," Scientific American, May 1997.

Forrest M. Mims III, "How to Monitor Ultraviolet Radiation From the Sun," Scientific American, August 1990.

Forrest M. Mims III, "The Sun & Sky Monitoring Station," reviewed by Sheldon Greaves.

Jearl Walker, "The Colors Seen in the Sky Offer Lessons in Optical Scattering,"Scientific American, January 1989.

Additional information on commercial particle detection and analysis can be found at:

Malvern Instruments (Requires Registration)and

Met-One Instruments

Breathe Pure Air

TSI Incorporated.

Part 3

In Part 3 we will look at ionization type smoke detectors. Their operation is quite different from that of the photoelectric types discussed here.

Contact Us | Privacy Statement | SAS Home Page
Copyright 2004 by Society for Amateur Scientists