Quick reference · Fire detection and alarm systems in non-domestic premises
BS 5839-1: quick reference
How to design, install, commission and maintain fire detection and fire alarm systems in buildings other than homes: where detectors and call points go, how loud the alarm must be, how the system is tested and serviced, and what is recorded.
The rulebook for fire alarms in shops, offices, factories and other non-domestic buildings: where to put detectors and call points, how loud the alarm has to be, and how the system is tested and looked after.
2025. BS 5839-1:2025 replaced the 2017 edition in April 2025.
Confirmed two or more sources agree, or a source quotes the standard. Check one source, or a detail that could not be confirmed. Check it against the standard.
Every point in a protected area should be within 7.5 m horizontally of a smoke or CO detector, or 5.3 m of a heat detector. Laid out on a square grid, that is about 10.6 m between smoke detectors and 7.5 m between heat detectors. The figures are unchanged from 2017.
A smoke detector looks after everything within 7.5 m of it. A heat detector looks after everything within 5.3 m. Space them so that nowhere is further away than that.
Nowhere should be more than 7.5 m from a smoke detector or 5.3 m from a heat detector. The dashed square is the biggest square one detector covers, which is what sets the spacing on a grid.
Point smoke and CO detectors: 10.5 m, or 12.5 m over up to 10% of the area. Heat detectors class A1: 9 m (10.5 m for 10%). Other heat classes: 7.5 m (10.5 m for 10%). A multi-sensor that works on heat and smoke only takes the heat limit.
Smoke detectors work up to 10.5 m high. The better class of heat detector works up to 9 m, and the others up to 7.5 m. Up to 10% of the area can be a bit higher.
The normal limit is the green part. Over up to 10% of the area a detector may go a little higher, up to the top of the brass part.
In a corridor under 2 m wide the coverage areas do not need to overlap, so smoke detectors can be about 15 m apart and heat detectors about 10.6 m. A corridor wider than 2 m is treated as a room.
In a corridor narrower than 2 m, detectors can be further apart: about 15 m for smoke and about 10.6 m for heat. Wider than 2 m, treat it like a room.
In a corridor under 2 m wide the coverage circles do not need to overlap: about 15 m apart for smoke and 10.6 m for heat. Wider than 2 m, space detectors as in a room.
A ceiling counts as flat if its apex is under 150 mm above the eaves for heat detection, or under 600 mm for smoke. Above that, a detector goes at or near the apex, and its radius alone can grow by 1% for each degree of roof pitch, up to 25%.
A slight slope is treated as flat. Steeper than that, put a detector at the top, and it can cover a little further out.
A slight slope counts as flat. On a steeper roof put a detector at the apex, and its radius alone can grow.
The wording is easy to misapply. Check the standard before relying on it.
At least 65 dB(A), or 5 dB(A) above any background noise that lasts more than 30 seconds, whichever is higher, with the doors closed. In sleeping areas the level is 75 dB(A) at the bed head. In stairways or enclosures under 60 m² (not corridors) it can be 60 dB(A).
Everywhere should hear the alarm at 65 dB(A) or more, or 5 louder than the background noise if that is higher. At a bed head it should be 75 dB(A). In small stairways and enclosures 60 is enough.
Measure with the doors closed. A normal door takes off about 20 dB(A) and a fire door about 30.
As a rough guide, a normal door cuts the sound by about 20 dB(A) and a fire door by about 30 dB(A). The real test is a measurement with the doors shut.
A normal door takes about 20 dB(A) off the sound and a fire door about 30. Always measure with the doors closed.
No more than 45 m along the route someone would walk. At design stage, when the layout is not known, use 30 m in a straight line. The reduced 25 m and 16 m distances in the 2017 edition were removed in 2025 (clause 19.5).
Nobody should have to walk more than 45 m to a break glass unit. If you are designing before the layout is fixed, use 30 m in a straight line.
No more than 45 m along the route walked. At design stage, when the layout is not known, use 30 m in a straight line.
1.4 m to the centre of the break glass, with a tolerance of 200 mm above and 300 mm below, so 1.1 m to 1.6 m (clause 19.8). Transparent hinged covers are recommended.
Fix them about 1.4 m from the floor. Anywhere from 1.1 m to 1.6 m is acceptable. A clear hinged cover is recommended.
1.4 m to the centre of the break glass, with 200 mm above and 300 mm below allowed: 1.1 m to 1.6 m.
L1 has detection throughout. L2 has defined areas plus the L3 areas. L3 covers escape routes and the rooms off them. L4 covers escape routes only. L5 is designed to a specific fire risk objective. M is manual call points only. P1 has detection throughout and P2 in defined areas, both to protect property. Add /M to P1, P2 or L5 for call points.
L is for life, P is for property, and M is call points only. The number says how much of the building has detectors: L1 all of it, L2 some areas plus the escape routes, L3 the escape routes and rooms off them, L4 just the escape routes.
Each plan has a corridor (the escape route), two rooms that open onto it, and two areas behind them. L is for life, P is for property, and M is call points only.
Category L2 now includes rooms where people sleep. The standard now recommends smoke, multi-sensor or CO detectors there, and heat detectors only if a fire risk assessment justifies them. The same applies to sleeping rooms in L3. It is not retrospective.
Rooms where people sleep are now covered by L2. Use smoke, multi-sensor or CO detectors there. Heat detectors need a risk assessment to justify them. Old systems do not have to change.
About every 6 months. A visit any time between 5 and 7 months after the previous one is acceptable (clause 43.2.1). The 2017 edition said the period should not exceed six months, and a late visit meant the system was no longer compliant.
Service it about every six months. A visit anywhere from 5 to 7 months after the last one is fine.
Service about every 6 months. A visit between 5 and 7 months after the last one is fine.
The panel clock is checked and adjusted at every visit, and alarm receiving centre signals are checked, confirming each signal where more than one can be sent. At the 12-month visit the zone identification on the panel is compared with the zone plan.
At every visit, check the panel clock, and test the signals to the monitoring centre. Once a year, check that the zone names on the panel match the zone plan.
Events, faults, false alarms, tests and disablements, and every variation from the standard that has been agreed. In the 2025 edition all variations must be recorded.
Write down every event, fault, false alarm, test and switch-off, and every agreed departure from the standard.
A departure from a recommendation in the standard, agreed with the person responsible and recorded, usually because following it exactly is not practical or does not suit the building. Some variations, such as leaving out a zone plan, are no longer acceptable in 2025.
A variation is an agreed exception to the standard, written down. Some exceptions are no longer allowed.
Annex E: Cmin = 1.25 × (T1 × I1 + D × I2 × T2), in amp hours at the 20-hour rate and 20 °C. 1.25 is the ageing factor and D is the de-rating factor, 1.75 unless the maker gives another figure. The usual defaults are 24 hours standby and 30 minutes alarm. The result is the same as under the 2017 method.
Work out the charge used in standby and in alarm, allow extra for the battery losing capacity with age, and choose the next size up. Use 24 hours standby and 30 minutes alarm unless told otherwise.
Worked example: 0.25 A standby for 24 hours, 0.75 A for 30 minutes of alarm, 25% added for ageing, then the next size up.
Annex G holds the model certificates: design, installation, commissioning, acceptance, verification, inspection and servicing, and extension or modification. A firmware update counts as a modification.
There is a certificate for each stage: design, installation, commissioning, handover acceptance, verification, each service visit, and any extension or change.
Guides describe detection in voids of 800 mm depth or more, in the top 125 mm for voids of 800 to 1,250 mm and in the top 10% for 1,250 to 1,500 mm. A void over 1,500 mm is treated as a room.
Deep voids may need detectors. The depth decides where they go. A very deep void counts as a room.
The deeper the void, the more it needs detection. A very deep void is treated as a room.
Within 90 seconds for Category L systems and 120 seconds for Category P. A complete transmission failure must show at the receiving centre and on the panel within 3 minutes for L, or 31 minutes for P.
The signal must arrive at the monitoring centre within 90 seconds (L) or 120 seconds (P). If the link fails completely, it must be noticed within 3 minutes (L) or 31 minutes (P).
Two clocks: how fast an alarm gets through, and how fast a broken link is noticed. The two bars use different scales.
A false alarm notice at or beside the panel, so people warn the receiving centre before a test. The centre should also be given premises information, such as whether people sleep there and, where practicable, the device type and whether coincidence filtering is used.
Put a label at the panel reminding people to tell the monitoring centre before testing. Tell the centre about the building, for example whether people sleep there.
All fire alarm cables and the low-voltage mains feed should be fire-resistant to the same grade, in one common colour, with red preferred. Black or white should not be used. 'Standard' fire-resistant cable means PH30 (BS EN 50200 with the Annex E water-spray test, 30 minutes) and 'enhanced' means PH120 (BS EN 50200 plus BS 8434-2, 120 minutes). Cable that is not fire-resistant but is mechanically protected no longer complies.
Use fire-resistant cable throughout, one colour, preferably red. Standard means 30 minutes of fire resistance and enhanced means 120 minutes. Ordinary cable in protection no longer counts.
Fire-resistant cable throughout, one colour, red preferred. PH30 is standard and PH120 is enhanced.
Yes. BS 5839-1:2025 (clause 24.3.3) recommends marking batteries with their installation date. Writing it on the casing in permanent marker is acceptable.
Write the date the battery was fitted on the battery, in permanent marker.
They are recommended in areas with high noise, in all sanitary accommodation, in hotel, student and other sleeping rooms, and wherever hearing-impaired people may be alone. They should be certified to BS EN 54-23 and are coded C-x-y (ceiling), W-x-y (wall) or O (open). A visual alarm device will not wake a sleeping person. This is guidance from the 2013 and 2017 era.
Flashing beacons are for noisy areas, toilets, bedrooms and anywhere a deaf person might be on their own. Use certified ones. A flashing light does not wake someone who is asleep.
The coverage rules come from 2013 and 2017 guidance. Check the 2025 standard.
Yes. The same sounder may carry non-fire signals using different tones. School class-change signals may last up to 10 seconds (it was 5). Remote indicators must show red and are checked at the annual service.
A fire sounder can also give other signals if the tone is different. School bells can be 10 seconds long. Remote indicators must be red and are checked each year.
Two. Leaving out a zone plan where a storey has more than one zone, and having no alarm receiving centre link in residential care or supported housing that needs Grade A.
You cannot agree to skip a zone plan on a floor with several zones, or skip the monitoring link in care homes that need a full system.
A cause-and-effect matrix or a plain-text description of what should happen, the detector type and settings chosen for multi-sensors, the zone plan and the log book, which may be digital or paper. Redundant devices are removed or labelled 'not in use'.
Hand over a description of what triggers what, the multi-sensor settings, the zone plan and the log book. Old devices should be taken out or marked as not in use.
The user should investigate every false alarm. The 2025 edition adds a new Annex F with a false alarm rate formula, and asks for a preliminary investigation when the rate is high.
Whoever looks after the building should look into every false alarm, and find out why.
The rate formula could not be confirmed. Check Annex F.
Someone with the training, knowledge and experience to do the work, who keeps that competence up, probably through continuing professional development. The designer signs a design certificate and ideally witnesses testing.
Someone who has the training and experience for the job and keeps their skills up to date.
Duct-mounted smoke detectors are tested. The zone shown on the panel is checked against the zone tested and the zone plan. Every detector, call point and sounder is tested within the year, and remote indicators are checked for obstruction or paint.
Once a year also test detectors in air ducts, and check that the zone names on the panel, the plan and what you tested all agree.
Lock communications cabinets, use anti-tamper patch plugs, and authenticate before any remote connection. A risk assessment is needed before remote read, control or write access.
Keep network cabinets locked, use tamper-proof plugs, and make people prove who they are before connecting remotely. Assess the risk before allowing remote changes.
If the gap above it is under 300 mm and its depth is more than 10% of the ceiling height. Keep detectors at least twice the depth away from an isolated obstruction up to 250 mm deep. Closely spaced beams are about 1 m apart or less.
A deep beam or partition close to the ceiling counts as a wall. Keep detectors well away from shallow obstructions, at least twice their depth.
A beam counts as a wall if the gap above it is under 300 mm and it is deeper than 10% of the ceiling height.
Guides give 25 m for beam detectors at normal sensitivity and 40 m at enhanced sensitivity, with supplementary detection unless stratification is unlikely. Aspirating detection of class B with at least 15 sampling holes reaches 40 m. Linear heat cable uses BS EN 54-22 (resettable) or 54-28 (non-resettable).
Beam detectors work up to 25 m, or 40 m with the more sensitive type. Special aspirating systems can reach 40 m too.
Guides say an enclosed stairway has a detector at the top and on each main landing, and open stairs, lift shafts and risers have one within 1.5 m of the opening. Stairways and shafts are separate zones.
On a closed staircase, a detector at the top and on each main landing. On open stairs and shafts, one within 1.5 m of the opening.
Enclosed stairs: a detector at the top and on each main landing. Open stairs, shafts and risers: one within 1.5 m of the opening.