Smoke detector causes fire

[Tom W]

Taken from info for fire
"An investigation is underway after a smoke alarm appeared to spontaneously combust in a home in Hull this week.

16-year-old Victoria Gray reported hearing a "large bang" which she thought was the family dog, Cookie. Upon further investigation she discovered a small fire in the utility room, which she extinguished herself with a bucket of water, according to the Daily Mail.

The Fire Angel ST620 was in a newspaper rack in the room at the time of the explosion. There have been reports that another fire was caused by the same model elsewhere in the UK last week.

The smoke alarms are supplied by fire and rescue services nationally as part of a programme to encourage fire safety. Humberside Fire and Rescue Service have decided to suspend providing the detectors while an independent laboratory investigates further.

A spokesman for Sprue Aegis, who produce the Fire Angel alarms, said that they are cooperating while the independent analysis is ongoing, and in the mean time it is "very much business as usual."

"Highly technical people cannot explain how it happened. We're not taking these reports lightly, however we provide one million alarms every year and we're talking about one small fire. This didn't happen in situ. The forensic side of it is looking into what happened to [the alarm] when it was taken off the ceiling."
 
"Highly technical people cannot explain how it happened" without wishing to blow any of your trumpets, anyone fancy stabbing a guess?

[Golden]

Its a difficult one! We're only looking for the heat source as the fuel is provided by the paper rack which probably also gave good insulation. The report I read was that the detector had stopped working (low battery signal) and the only thing I can assume is that it was an internal fault rather than a low battery - again I believe it was a sealed unit so poor connection of the battery can also be ruled out.

So my conclusion from about 300 miles away is that the detector had developed a fault which was generating a small amount of heat; when the detector was removed from the ceiling and put into the paper rack the heat couldn't escape and ended up generating sufficient heat to ignite (did it ignite or was it smouldering?) the paper. An unusual cause with a very low probability but I'd imagine its plausible. 

This is also assuming it was the detector that caused the fire ... otherwise careless disposal!!

[Dinnertime Dave]

Around half a million of these alarms have been supplied to FRS throughout the country.

One thought, many years ago I remember TV remote controls causing fires when they had fallen down the back of sofas, if my memory serves me right the buttons were held in creating heat - Something similar?   

[kurnal]

Some battery systems can be unstable - I have no idea what type of battery is used in these things but for example Garmin recalled sat navs fitted with Lithium batteries due to a fire risk, and Sony had a similar issue with laptops last year.

[SamFIRT]

I believe it was a sealed unit so poor connection of the battery can also be ruled out.

Why's that Golden ?  

It is entirely possible for factory connections to be faulty, especially if a problem is associated with a batch, where there many have been failure of quality control

Time will tell. I for one would like to be investigating this problem.  

Oh and there are two threads on this topic on the forum; one in Community Safety and this one. Should it not be in Fire Investigation instead of these two locations?

Oh PS

did it ignite or was it smouldering?)

Smouldering is ignition. You can have smouldering ignition or flaming ignition. One can lead to the other and or back again. Char however is not ignition.

[Golden]

I was being pedantic on another thread and should have continued it to this one ... sorry for not taking long enough to explain that I meant that the user had not connected the battery properly as I'd assumed that as it had worked for a number of years that the connection would be OK from the factory. Also that I meant without flaming ignition; although smouldering is defined in most dictionaries as "burn slowly with smoke but no flame" which is kinda what I meant.

I never wanted to be an FI and now I know I would never have made it work - back to the day job.

[SamFIRT]

I believe in this case it is being said all the detectors were fitted in September, which implies they have not been in use for long and they may therefore have come from the same batch. An area for investigation and research.

Also it is entirely possible for a high resistance fault to become progressively worse over a very long period, heating exponentially as a semiconducting circuit (or tracking) is created in the resultant char and thereby flattening the battery as energy is progressively consumed creating the ever increasing heat.

As I said an interesting investigation.