Initiating circuits and supervision

Cartwright Distribution Centre has more supervisory points than alarm points. Its sprinkler system has four risers, each with a control valve, a waterflow switch and a pressure gauge; the dock has a dry-pipe system with a low-air switch; the fire pump has running, power-failure and phase-reversal signals; and the heated valve rooms have low-temperature switches. Every one of those is a contact device wired to an input module, and every one has to report the right kind of signal at the right time. Gerry walks Sasha down the riser room wall. 'The detectors get all the attention,' he says, 'but in a sprinklered warehouse, these switches are the fire alarm. If the waterflow switch doesn't report, nobody knows the sprinklers are fighting a fire until the smoke reaches the office.' This lesson covers how the panel monitors contact devices, how it supervises their wiring and what it does with each kind of signal.

Initiating device circuits

An initiating device circuit is any circuit that brings a contact or a two-wire detector's signal into the panel without the device having its own address. It may be a conventional zone on the panel itself, or the short circuit between an addressable input module and the switch it watches. FAT-130 covers the electrical side: the supervisory current, the end-of-line resistor and the voltage drop. Here the focus is on what the panel does with the circuit and how it is wired so that every fault is caught.

The rule that matters most is that the panel must be able to detect the removal of any device or the breaking of any wire. That is why the wires must be broken at each device terminal, with the incoming wire on one terminal and the outgoing wire on another, or with separate incoming and outgoing conductors under the same terminal where the device is listed for it. Looping one continuous wire under a terminal screw without cutting it means that if the device is removed, the circuit stays complete and the panel never knows. The installation standards in both countries forbid this, and an inspector who finds an unbroken loop will reject the circuit.

The end-of-line device must be at the electrical end of the circuit, after the last device, so the supervisory current passes every connection. Putting the end-of-line resistor at the panel or at the module, as some installers do for convenience, leaves the whole field wiring unsupervised. Where a module is mounted beside the switch it watches, the end-of-line device goes in the switch housing or the box beside it, as the module's manual shows. Some modules have an end-of-line resistor built in for exactly this case, and the manual says so.

What the panel sees

On a supervised contact circuit the panel recognizes four basic states, and the module or zone reports each one. Normal means the supervisory current matches the end-of-line value. Active means a device has closed its contacts, raising the current; the panel reports this as alarm, supervisory or another signal depending on how the point is programmed. Open means the current has dropped to near zero because a wire has broken or a device has been removed; the panel reports a trouble. Short, on some circuits, means the two wires have touched; depending on the circuit and its programming it is reported as a trouble or treated as active. Many panels also measure the circuit's leakage to ground, which the next section covers.

# Decoding an input module circuit from its measured resistance (illustrative bands)
R_EOL = 47000            # end-of-line resistor value (ohms) for this module
bands = [                # (upper limit in ohms, state reported)
    (100, "short"),
    (20000, "active"),
    (80000, "normal"),
    (10 ** 9, "open"),
]
def state(r):
    for limit, name in bands:
        if r < limit:
            return name
readings = {"WF-1 riser 1": 46800, "WF-2 riser 2": 2200, "TS-3 riser 3 valve": 10 ** 8,
            "LA-4 dock low air": 47100, "LT-5 valve room cold": 35}
for point, r in readings.items():
    print(f"{point:22s} {r:>11,} ohms -> {state(r)}")

WF-1 riser 1                46,800 ohms -> normal
WF-2 riser 2                 2,200 ohms -> active
TS-3 riser 3 valve     100,000,000 ohms -> open
LA-4 dock low air           47,100 ohms -> normal
LT-5 valve room cold            35 ohms -> short

The output reads like a quick tour of the riser room: riser 2's waterflow switch is active, riser 3's tamper circuit is open and needs fixing, and the valve room low-temperature circuit is shorted. The bands are illustrative; each module's manual gives its own end-of-line value and the ranges it uses.

Ground faults

A ground fault on a fire alarm circuit is a conductor that has made an electrical connection to the building's metal or to earth, through damaged insulation, a wire pinched against a box, or water in a junction box. One ground fault usually doesn't stop the circuit working, but two on different conductors can create a path that looks like a short or an alarm, or that disables a whole section. For that reason panels monitor every field conductor for leakage to ground and report a ground fault trouble when it falls below a threshold. Most panels report one common ground fault for the whole system, or one per power supply, so finding which circuit it is on is a technician's task.

Finding a ground fault starts at the panel. With the building informed and the system watched, the technician disconnects field circuits one at a time, checking after each whether the ground fault trouble clears. When it clears, that circuit holds the fault; the circuit is then halved with a meter set to measure resistance from each conductor to ground. Some panels show the ground fault per loop or per circuit, which saves time. A common cause in warehouses like Cartwright is condensation in junction boxes on outside walls or in unheated areas, and the fix there is a sealed box and a drip loop, not just new wire.

Supervisory signals and how they behave

Supervisory devices watch the sprinkler and suppression systems, and each type behaves a little differently. A valve tamper switch operates when a control valve moves away from fully open, within a small fraction of its travel, so that a valve being closed for maintenance or tampered with is reported long before it is fully shut. A low-air switch on a dry-pipe system operates when the air pressure falls below its setting, warning that the system may trip or freeze. A low-temperature switch in a valve room operates before the water in the pipes can freeze. Fire pump signals report the pump running, a loss of power or a phase reversal. Each is programmed as a supervisory point, not an alarm.

Supervisory signals are latching or non-latching depending on the device and the code. A valve tamper is usually non-latching, so it restores when the valve is reopened, which tells the operator the valve is open again. A low-air or low-temperature signal latches until reset on some systems, so a brief condition isn't missed. A fire pump running signal is a supervisory signal on most systems; in some jurisdictions a pump running is treated as an alarm. The sequence of operation, covered in lesson 9, says which behaviour each point needs, and the panel is programmed to match.

At Cartwright, the sprinkler fitters close valves several times a year for maintenance. The building's procedure, required by the fire code and the inspection standard, is that the fire alarm contractor or the monitoring centre is told first, the closed valve is recorded as an impairment, and the panel's supervisory signal is acknowledged but never disabled. A supervisory point disabled for a valve closure and forgotten is how sprinkler systems end up shut off for months.

Waterflow and the retard delay

A waterflow switch detects water moving in the sprinkler pipe, which normally means a sprinkler head has opened. But water also moves briefly when the city's water pressure surges, a pump starts, or air trapped in the pipe compresses. To avoid false alarms from these surges, most vane-type waterflow switches have a mechanical or electronic retard, a delay that the flow must outlast before the switch closes its contacts. The retard is set on the switch, and the panel should not add a second delay on top of it, because the total delay before the alarm would then be too long.

The standards limit the total time from the start of flow to the alarm signal, commonly to 90 seconds, measured from the moment water starts to flow from the inspector's test connection. During the inspection, the sprinkler fitter opens the test valve and the technician times how long the panel takes to show the alarm. A switch whose retard has been set to its maximum, or a panel input with an added delay, can push the total over the limit. Riser 3 at Cartwright needed its retard reduced from the factory setting after the last test for exactly this reason.

# Waterflow test timing: start of flow to panel alarm (limit 90 s)
LIMIT = 90
tests = [   # (riser, switch retard setting in s, extra delay programmed at the panel in s, measured total in s)
    ("Riser 1", 45, 0, 52),
    ("Riser 2", 60, 0, 71),
    ("Riser 3", 90, 0, 104),
    ("Riser 4", 30, 30, 69),
]
for riser, retard, panel_delay, total in tests:
    issues = []
    if total > LIMIT:
        issues.append(f"over the {LIMIT} s limit")
    if panel_delay:
        issues.append("delay added at the panel; remove it")
    print(f"{riser}: retard {retard:3d} s, panel {panel_delay:2d} s, measured {total:3d} s -> {'; '.join(issues) or 'pass'}")

Riser 1: retard  45 s, panel  0 s, measured  52 s -> pass
Riser 2: retard  60 s, panel  0 s, measured  71 s -> pass
Riser 3: retard  90 s, panel  0 s, measured 104 s -> over the 90 s limit
Riser 4: retard  30 s, panel 30 s, measured  69 s -> delay added at the panel; remove it

Non-fire inputs and their place

Fire alarm panels often monitor things that aren't fire signals: a kitchen hood suppression system, a clean agent system, a generator running signal, or a door that is supposed to stay closed. These inputs are programmed according to what they are. A kitchen suppression system discharging is normally an alarm input, because it means there is a fire. A generator running is a status signal with no fire response. The panel must keep fire signals ahead of everything else in priority, and the standards limit how non-fire signals may be combined with fire alarm functions so that they can never delay or hide a fire signal.

Supervising the power to control devices

Some devices on initiating and control circuits need their own power to work, such as a four-wire smoke detector, a duct detector with its own relay, or a beam detector, and that power must be supervised too. The usual method is an end-of-line power supervision relay at the last powered device: the relay is held energized by the power circuit, and its contact is wired in series with the initiating circuit. If the power fails anywhere along its run, the relay drops out, the initiating circuit opens, and the panel reports a trouble. Without that relay, a lost power feed would leave the devices dead with the panel showing normal.