Article on domestic smoke alarms- International Fire Risk Professional Journal

[kurnal]

An interesting article in the Journal comparing the relative performance of ionisation and optical smoke alarms in the event of smouldering fires. The findings were not surprising other than a summary dismissive comment that CO detectors failed to operate, It did not clarify if these were CO fire alarms or CO detectors.

The report also suggests that for high risk situations wall mounted smoke alarms may give earlier warning of smouldering fires than ceiling mounted alarms, though again it implies that the alarms would need to be mounted on the wall nearest to the fire.

What is really needed is a study on the performance of smoke alarms vs tenability criteria to identify whether a smoke detector is capable of giving an alarm of fire in time to save the life of the occupant of the room, rather than as a means of protecting the escape routes outside.

[John Webb]

The erratic response of domestic detectors to smouldering fires is not a new one. Work I did for the Home Office back in 1991 (corrected date) and published in summary form by their Fire Research and Development Group ("Comparison of British Standards BS 5445 part 7 and BS 5446 part 1" by C A Reynolds) showed a similar trend. One thing they don't seem to have done, and which I did, was to look at what the detectors were seeing by connecting up measuring equipment to their sensor circuits. It was these measurements which indicated that smoke was not getting through the light traps needed in optical detectors; we (that is the HO and myself) concluded that the smoke velocity from the various smouldering fires was too low to work its way round the labyrinth and into the sensing area. The ionisation detector has a more open arrangement which makes the low velocity less of a problem.

This may also explain the better operation on a wall - the smoke and heat is rising up the wall in possibly a more concentrated form and might have a higher velocity than when it reaches the ceiling and starts spreading out.

I don't have a copy of the current BS EN 54-7:2001, and so I'm uncertain exactly how the test fires in this compare with the earlier two standards (as a reminder BS 5445 pt 7 was for detectors in commercial and industrial alarm systems while BS 5446 pt 1 was for domestic detectors).

But one of my criticisms of the test fires in BS 5446 was the degree to which they were (or were not) representative of a fire in domestic premises. I suggested that tests should be done to measure air velocity and other parameters from typical fires in domestic premises. As far as I am aware, no work was subsequently done on this.

In 1983 less elaborate tests were conducted on a range of domestic detectors in the three-bedroom two storey house in the hanger at Cardington, also for the Home Office, on the response of such detectors to smouldering fires in upholstery. I'm uncertain if the HO published these results and in what format. Again there were differences in the behaviour of the two types of detector, also dependent upon being in the room of origin or outside it, and if the door of the room of origin was closed or open.

It would be interesting to know if the authors of the present article were aware of the above works or not.
 

[Mr. P]

BS EN 54-7 2001 page 9

4.8 Response to slowly developing fires
The provision of ?drift compensation? (e.g. to compensate for sensor drift due to the build-up of dirt in the
detector), shall not lead to a significant reduction in the detector?s sensitivity to slowly developing fires.
Since it is not practical to make tests with very slow increases in smoke density, an assessment of the
detector?s response to slow increases in smoke density shall be made by analysis of the circuit/software,
and/or physical tests and simulations.
The detector shall be deemed to meet the requirements of this clause if this assessment shows that:
a) for any rate of increase in smoke density R, which is greater than A/4 per hour (where A is the
detector?s initial uncompensated response threshold value), the time for the detector to give an
alarm does not exceed 1,6 ? A/R by more than 100 s; and
b) the range of compensation is limited such that, throughout this range, the compensation does not
cause the response threshold value of the detector to exceed its initial value by a factor
greater than 1,6.

[Mr. P]

Sorry - here's another bit - if you want more, let me know.

EN 54-7:2001
? BSI 2006
5 Tests
5.1 General
5.1.1 Atmospheric conditions for tests
Unless otherwise stated in a test procedure, the testing shall be carried out after the test specimen has been
allowed to stabilize in the standard atmospheric conditions for testing as described in
IEC 60068-1:1988+A1:1992 as follows:
a) temperature: (15 to 35) ?C;
b) relative humidity: (25 to 75) %;
c) air pressure: (86 to 106) kPa.
NOTE: If variations in these parameters have a significant effect on a measurement, then such variations should
be kept to a minimum during a series of measurements carried out as part of one test on one specimen.
5.1.2 Operating conditions for tests
If a test method requires a specimen to be operational, then the specimen shall be connected to suitable
supply and monitoring equipment with characteristics as required by the manufacturer?s data. Unless
otherwise specified in the test method, the supply parameters applied to the specimen shall be set within the
manufacturer?s specified range(s) and shall remain substantially constant throughout the tests. The value
chosen for each parameter shall normally be the nominal value, or the mean of the specified range. If a
test procedure requires a specimen to be monitored to detect any alarm or fault signals, then connections
shall be made to any necessary ancillary devices (e.g. through wiring to an end-of-line device for
conventional detectors) to allow a fault signal to be recognized.
NOTE: The details of the supply and monitoring equipment and the alarm criteria used should be given in the
test report

[John Webb]

Does the BS EN 54 still have provision for testing detectors in a wind-tunnel for response thresholds prior to fire tests, please?

[wee brian]

This Article sounds worth a read - is it on-line anywhere.

[Mr. P]

Annex A
(normative)
Smoke tunnel for response threshold value measurements
The following specifies those properties of the smoke tunnel which are of primary importance for making
repeatable and reproducible measurements of response threshold values of smoke detectors. However,
since it is not practical to specify and measure all parameters which can influence the measurements, the
background information in annex K should be carefully considered and taken into account when a smoke
tunnel is designed and used to make measurements in accordance with this standard.
The smoke tunnel shall have a horizontal working section containing a working volume. The working volume
is a defined part of the working section where the air temperature and air flow are within the required test
conditions. Conformance with this requirement shall be regularly verified under static conditions, by
measurements at an adequate number of points distributed within and on the imaginary boundaries of the
working volume. The working volume shall be large enough to fully enclose the detector to be tested and the
sensing parts of the measuring equipment. The working section shall be designed to allow the dazzling
apparatus described in annex D to be inserted. The detector to be tested shall be mounted in its normal
operating position on the underside of a flat board aligned with the airflow in the working volume. The board
shall be of such dimensions that the edge(s) of the board are at least 20 mm from any part of the detector.
The detector mounting arrangement shall not unduly obstruct the air flow between the board and the tunnel
ceiling.
Means shall be provided for creating an essentially laminar air flow at the required velocities
(i.e. (0,2 ? 0,04) m s-1 or (1,0 ? 0,2) m s-1) through the working volume. It shall be possible to control the
temperature at the required values and to increase the temperature at a rate not exceeding 1 K min-1
to 55 ?C.
Both aerosol density measurements, m and y, shall be made in the working volume in the proximity of the
detector.
Means shall be provided for the introduction of the test aerosol such that a homogeneous aerosol density is
obtained in the working volume.
Only one detector shall be mounted in the tunnel, unless it has been demonstrated that measurements made
simultaneously on more than one detector are in close agreement with measurements made by testing
detectors individually. In the event of a dispute the value obtained by individual testing shall be accepted.

[John Webb]

Thanks, Mr P. As I recall, the wind-tunnel construction was quite precisely detailed in the older BSs to try and get similar test conditions in different test labs. I must have tested a couple of hundred detectors or more in the tunnel before we handed over that work to the Fire Insurers Research and Testing Organisation in the mid-1970s.