There are two kinds of smoke detector on the market, but one of them doesn’t deserve the name. Paul Skelton reports.

Ionisation alarms were introduced in 1972 and now feature in about 90% of Australian homes.

The trouble is, they were never intended to be smoke detectors ­– they are designed to detect flame. By installing such a unit you could be putting your customers’ lives at risk.

When photoelectric smoke detectors entered the market they proved to be a capable replacement for ionisation alarms in homes. It seemed like an easy path ahead for photoelectrics, but that was not to be.

What’s the difference?

Ionisation and photoelectric smoke alarms are designed to detect fire, but the different technologies mean they detect fires at different stages.

Ionisation smoke alarms ‘smell’ the smoke that comes from the flames of a fire. The ionisation chamber is charged with electrical particles, ions, by a small amount of radioactive material.

Unlike the photoelectric model, ionisation alarms are sensitive to small particles of combustion that disrupt the balance of ions, causing the alarm to sound. This means they are marginally better at detecting flaming fires that produce smaller amounts of visible smoke.

Photoelectric units ‘see’ the smoke before it bursts into flames. The chamber has a light source projected into it, so when visible smoke enters the chamber it scatters and disturbs the light. The alarm sounds once the large smouldering fire particles are detected by the light-sensitive receiver.

Both types are effective in detecting most types of fire, but photoelectric alarms are more advanced at detecting smouldering fires, which are most often associated with residential fatalities.

According to the Fire Protection Association of Australia, this is because smouldering fires are more likely to occur while occupants are sleeping and won’t have enough time to avoid the fire’s effects.

On the other hand, ionisation alarms are marginally faster at detecting fast flaming fires.

Flaming fires are still a threat, but they are most likely to occur when residents are alert – stoves or ovens are sources of high heat, and gas cooktops introduce naked flame.

It’s also important to note that ionisation models are prone to false alarms, as the technology can be set off by high humidity from showers, harmless cooking steam or smoke from burning toast. This is important. A high incidence of false alarms leads to occupant complacency and is an incentive to disconnect the alarm.

The ionisation problem

It seems simple: photoelectric models are seemingly far superior to ionisation for residential applications, and yet ionisation alarms still dominate the market.

Adrian Butler is a retired firefighter who started The World Fire Safety Foundation in 2000 with the goal of educating the market about flawed detectors and to have ionisation alarms eliminated from Australian homes.

Adrian estimates that as many as 90% of homes have ionisation units installed, and they will do nothing in the event of a fire.

“This issue can be traced right back to the 1970s. The International Association of Fire Chiefs commissioned a report that found that ionisation alarms weren’t safe, lives were at risk and fire chiefs should recommend only photoelectric alarms,” he says.

“The battle has raged for more than 40 years.

“Perhaps foolishly, I believed that when we started The World Fire Safety Foundation in 2000 the problem would be resolved in a couple of years. The evidence back then was so profound.”

And yet very little has changed – until recently.

Before starting the foundation, Adrian ran a successful franchise in Australia and New Zealand that sold ionisation alarms.

“We sold between 10,000 and 20,000,” he says.

“But then we noted that these alarms would go off when people were cooking but not when there was an actual fire.

“Some people just refuse to believe there is an issue. Their alarms go off when they burn their toast, so they think their alarm is really effective. It isn’t.

“Ionisation alarms are sub-micron particle detectors. They will detect sub-micron particles emitted by the heating elements of a toaster or griller. However, they won’t detect smoke, even though they are sold with ‘smoke detector’ on the packaging.

“In the middle of the night, if your house catches fire – specifically a smouldering fire, the type that kills – ionisation alarms remain silent.

“Until there’s sufficient heat to generate sub-micron particles they don’t make a sound.”

Adrian says a house can fill with deadly smoke, and it’s not until the flames erupt that an ionisation alarm will activate.

At that point it’s usually too late for a safe escape.

“The solution isn’t actually that difficult. The industry just seems to want to make it complicated so they don’t have to admit that they screwed up.

“Ionisation smoke alarms are so dangerous they should be banned and recalled.”

If Adrian is correct that up to 90% of installed alarms in the country are ionisation units, the question must be asked: ‘who is benefitting?’

“I think it’s more a case of manufacturers being afraid of litigation if they have to admit that the alarms they have sold for decades don’t work,” he says.

Testing times

In order for a smoke detector to be sold in Australia, it must be tested by CSIRO to ensure it complies with Australian Standard 3786.

Section 2.1 of the Standard states “smoke alarms shall be designed to respond reliably to the presence of smoke”. Adrian says this is the primary reason why ionisation alarms should be outlawed.

“As ionisation detectors don’t respond to smoke, obviously compliance with this part of the Standard is not possible.

“However, some manufacturers decided to change the definition of the word ‘smoke’ from visible particles of combustion to include invisible particles. Effectively, they have rewritten the dictionary.

“So, while the public perception of smoke is that it is something you can see, the industry disagrees. And that’s where all the problems arise. Sub-micron particles aren’t present in sufficient numbers to set off an ionisation alarm in the early, smouldering stage of a fire.”

The main test involves what is called ‘light obscuration’, which measures the level of smoke that must be present to activate an alarm. Photoelectric alarms, for example, must go off before light obscuration reaches 15%.

“As for ionisation alarms … well, nobody knows exactly,” Adrian says.

“Whenever we have asked for testing data from the manufacturers or from CSIRO, which performs the compliance tests, we are told it’s commercial in confidence. But somehow this is only the case for ionisation alarms and not photoelectrics.

“I even got my local member of parliament to ask for the data, and they refused to tell him. Why?”

Recently, Adrian was asked to testify as an expert witness in a court case in Alabama after three young girls perished in a house fire. The home was fitted with three operational ionisation alarms.

“We told the family’s lawyers that the manufacturer, which sells its products in Australia, needed to produce its testing data, to show the level of smoke at which its ionisation alarms activated. Within 48 hours of the data being received, the case was settled out of court, a confidentiality order was placed on the settlement and a protection order was placed on the documentation.

“Now, do you think that if ionisation alarms activated at a safe level the manufacturer would have done this?”

In 2007 the Standards Committee FP-002 Fire Detection, Warning, Control and Intercom Systems identified an issue in the testing of ionisation alarms and tried to have an anomaly fixed.

“Specifically, the current edition allows two pass criteria for the same product [smoke alarms], resulting in different performance outcomes,” the group wrote.

“Australia is the only country that uses two different pass criteria: all other regional and international Standards use an acceptance criterion based on light obscuration.

“CSIRO has reported to FP-002 that the different criteria result in significant differences in the performance of smoke alarms.

“Photoelectric smoke alarms, when tested in accordance with the requirements of AS 3786 typically respond between 8% and 16% obscuration per metre (Obs/m) while ionisation smoke alarms typically respond between 40% and 60% light Obs/m, with the majority of ionisation smoke alarms operating at the least sensitive end of this range.

“Under the current Standard, ionisation smoke alarms are permitted to have a lesser response to obscuration, which results in a significant negative impact on the Available Safe Evacuation Time.

“Australian and international research demonstrates that the highest number of fatalities in residential fires occurs between 8pm and 8am when occupants are typically sleeping, and these fires typically begin with a smouldering phase. Of principle concern is the impact of resultant smoke obscuration and toxic species on the occupants’ ability to escape.”

Ultimately, the group was unsuccessful in changing the Standard to make testing consistent across the two types of detectors.

Dismissing evidence

Despite the wealth of evidence against ionisation alarms, some groups are staunchly against changing the Australian Standard or the National Construction Code to favour photoelectrics in residential applications.

The reasoning put forward is sometimes unclear.

CSIRO executive director of manufacturing, digital productivity and services Anita J. Hill made a submission to an inquiry into the use of smoke alarms for preventing fatalities caused by smoke and fire.

Dr Hill’s submission to the Senate Legal and Constitutional Affairs Committee included:

“CSIRO is aware of a longstanding industry debate around smoke detection technologies used in alarms, usually categorised as photoelectric or ionisation types …

“CSIRO’s review of the results of Australian and international research indicates a number of viewpoints on the merits of each detection technology.”

In an interview with 60 Minutes the general manager of the Australian Building Codes Board (ABCB), Neil Savery, said: “There’s very little difference [between the two detector types] at the end of the day in terms of adequate warning. The board is not satisfied at this stage that it needs to make any further changes to the code.

“It hasn’t been demonstrated to the board’s satisfaction that one performs better in all circumstances over the other and the board doesn’t delineate between different types of products.

“If there was an overwhelming body of evidence that that was the case then the board would give that serious consideration, but that won’t be reflected at the mandating of a particular product.”

Interestingly, and perhaps a little confusingly, photoelectric smoke detectors have been mandated in commercial buildings since May 2004, but the ABCB has been outwardly dismissive of making the same ruling for the residential market.

Adrian says he just doesn’t understand this position.

“Why would the ABCB pass legislation to mandate the installation of photoelectric smoke detectors in all commercial buildings but not for homes?

“The government has even released a flyer called Wake Up To Photoelectric Smoke Alarms, which tells people they should replace ionisation alarms. But the ABCB still somehow refuses to make a change to the Standard.”

A changing tide

The issue first made national headlines in 2004 after A Current Affair, then hosted by Ray Martin, ran an exposé on the technology.

Twelve years later and finally it seems that people are listening.

In August, the Queensland Government passed legislation mandating photoelectric alarms.

Every Queensland residence will need to be fitted with photoelectric, interconnected smoke alarms in all bedrooms and hallways of dwellings.

The Minister for Police, Fire and Emergency Services, Bill Byrne, says it won’t matter which part of a house catches fire, the alarm closest to the occupants will sound.

“If you are asleep, an alarm will sound in your room, even if the area is closed off to the rest of the house.

“Research shows that photoelectric, interconnected smoke alarms are the most effective on the market for alerting people early.”

Mr Byrne says a 10-year phased rollout of the legislation would allow ample time for everyone to have alarms installed correctly.

“Hard-wired, interconnected photoelectric smoke alarms will require a qualified electrician to conduct the installation and ensure that alarms are working as they should.

“There is also an option to install photoelectric alarms with a 10-year lithium battery that have the capability to achieve interconnectedness wirelessly. This option may be more suitable for Queenslanders living in remote areas where attendance of an electrician could be difficult.”

All houses being built or substantially renovated must comply with the smoke alarm legislation upon completion after 1 January 2017. All houses leased or sold must comply after five years, and all owner-occupied private dwellings must comply within 10 years. Any smoke alarm replacement after 1 January 2017 must involve a photoelectric unit.

After 17 years of fighting to have ionisation alarms banned, Adrian says he is delighted by Queensland’s new law.

“People tend to think I’m a zealot and don’t know what I’m talking about. Everything I say can be backed up with documented evidence.

“I don’t sell smoke alarms, and I don’t solicit or accept donations. And now I just want this whole thing to come to a head because it has gone on far too long. Too many people have died, and I don’t want to keep on doing this campaign. It needs to be over.

“Actually, it should have been over 40 years ago.

“The scary thing is that most of the people reading this article will go to go to bed with a smoke alarm on their ceiling that is nothing more than an ornament.”