Smart alarm management
A signal arrives, and the next few seconds decide everything after them. The queue puts it where its consequence belongs rather than where its timestamp does. The site opens before the sound stops. Groups go out with the drawings, the photographs and the written instruction for this site in front of them. And when the outcome was never in doubt, the event closes itself without costing an operator anything.
The lifecycle
Five states, and an alarm is always in exactly one of them.
New
The event lands in the unhandled list and the audible alert starts. If the operator was looking at something else, the desk switches to the list by itself after a set delay. Until somebody takes it, the row stays marked as untaken.
Accepted
One action takes it. The sound stops, the row clears, the site card opens, and from here the event carries the name of the operator handling it.
Groups dispatched
Response groups are chosen from the ones actually free, with the written instruction for this site beside them. More can be added while the first is still moving, and a group can be recalled without closing the event.
Groups on site
Arrival is marked, and the event says so before anyone has to ask over the radio. A group needed somewhere else is released from here rather than by ending the event early.
Handled
Handling ends with a cause taken from the centre’s own list. The event leaves the queue, every group returns to free, and the account of who did what, and when, is already written.
These are not a recommended workflow. They are what the software will let happen: an event cannot be closed without a cause, and a cause cannot be invented at the desk.
Arrival
The queue orders itself by consequence, not by clock.
Panic buttons
Somebody has pressed something. First, always, whatever else is waiting.
Fire
Then fire, ahead of every intrusion signal in the list.
Other alarms
Zone violations, tamper, opening under duress.
Everything else
Missed tests, mains loss, a panel gone quiet. Real work, but not while something above it is open.
- A second, lesser alarm from a site that already has one goes to the archive and stays out of the queue until the first is closed. One site, one thing in front of the operator.
- Rows nobody has taken are marked as such, so a glance across the list tells you what is unclaimed rather than what is unread.
- A position can show only its own workload or every unhandled event in the centre — which is the difference between an operator and a supervisor.
- Queue and archive filter on the same fields: site, partition, zone, code, date and time. Clear the fields and you are looking at everything again.
The desk
Everything about the site is one action from the row, not one search away.
On the rail
Card
The whole site record, read at the desk in the same shape the administrator authored it.
State
What is installed and how it is doing: mains present, when it last tested, signal level, the firmware actually running, and any fault standing against the site right now.
Drawings
Approach routes and floor plans for the partition, on screen before the crew reaches the street.
Photographs
What the entrance looks like, so a crew arriving at night recognises it.
Test bench
Sites under commissioning, which technician is on them, and until when.
Response groups
Every group, its state and where it is — with the technicians beside them, showing who is working on what.
Stop list
Sites whose events are being discarded deliberately, and the period each one is off the line for.
Dispatch log
Every dispatch the centre has made, by stage, by operator and by the reason it was closed.
Remote control
The commands this panel actually supports, on the channel it is actually using, and nothing it does not.
Print
Whatever is in front of you, to paper or to a file, without leaving the position.
Signing off is a button, not a habit. The position returns to the sign-in screen without closing anything, so the next operator starts under their own name and the log knows where one shift ended.
Before the alarm
Most of the desk’s work is the alarm that has not happened yet.
Not armed on time
Sites that should be closed by now and are not — separated into the ones the contract says to phone, the ones the schedule expects, and the ones whose schedule nobody has set yet.
No test on time
Panels that have gone quiet, counted per channel so a dead SIM is not read as a dead site. The count restarts on a power-up, and a silent data channel waits for the voice channel before either is called a failure.
Faults
Zones opened and never restored, mains or standby power missing, expanders out of contact. The three things that become a false alarm at three in the morning if nobody sends a technician today.
Arming ban
Sites that may not be set at all, kept as a list rather than as something an operator is supposed to remember.
By status
Set and unset, with the time the status changed and the schedule the site was meant to be following.
External desks
Sites belonging to another centre whose alarms your groups answer, in the same list as your own.
Operator notes
An operator can leave a note on a site in any of these lists, with a date it stops being true. A site carrying a live note is marked one way in the list and an expired one another, so the next shift can see at a glance which notes still mean something. The note belongs to the list rather than to the site: when the site stops qualifying — the fault cleared, the test came in — the note goes with it.
Noise
An alarm that was never going to be real should not cost an operator anything.
- An ordinary siteEverything it sends reaches the operator.
- On the test benchCleared without an operator, kept in the archive.
- On the stop listNot accepted, not answered, not recorded.
Test bench
A site, a partition or a named zone under works is set aside for a stated period, or until the site is unset, or until the technician marks themselves on site. Everything from it clears automatically — and the fact that it was set aside, by whom and for how long, is in the archive.
Stop list
A panel flooding the centre through a fault is taken off the line for a fixed period. The centre sets its own thresholds for doing that by itself, and the desk can add or release one by hand.
Ignored
Signals you have decided you never want to see: dropped before the archive, though the panel goes on sending them.
Withdrawn keys
A lost or stolen key still works on the panel. It does not work here — setting or unsetting with one raises an alarm instead. The desk can add a key to the list; only an administrator can take one off it.
Handled on unset
When a site begins to unset, alarms from it are held for one minute. If the unset completes, they close themselves with the outcome the centre configured. If it does not, they come back as new and the operator has lost nothing.
Batch handling
Non-alarm events of one kind clear together instead of one row at a time. Alarms are deliberately excluded — those are handled one by one or not at all.
Seeing it
Two things an operator should never have to describe over the phone.
On the map
A site with coordinates opens on the map with the alarm marked on it, and groups are chosen and sent from the map itself rather than from the list. For a moving target the position keeps updating while the decision is being made.
On camera
Cameras are attached to zones, not to sites, so the alarm opens the camera covering the detector that fired. If no camera is assigned to that zone the window says so rather than showing the wrong picture. Stills and clips are saved from the same place.
The button that stays pressed
A panic button raised from a phone stays latched on the sender’s screen until the desk finishes handling the event. Closing it releases the button automatically, so the person holding the phone can raise the next one.
In the field
The dispatch reaches the vehicle, not the radio.
Desk to vehicleOne hop out, and what each crew can see once it is moving.
It arrives, and it is acknowledged
The tablet sounds and shows the target the moment the operator sends the group — a few seconds on a working mobile network. The crew accepts or declines, and the desk sees which. Every later change of state on the target follows it out.
Everything the desk has
Where the target is, what it is, its drawings and its photographs — in the van rather than in a folder at the office. A crew with no active target can still look a site up by number.
Maps that work without coverage
Online maps where there is a network and offline sets prepared in advance where there is not, so a black spot does not take the map away at the moment it is needed.
Crews that can see each other
Send more than one group and each of them sees the others as well as the target. On a moving target — a stolen vehicle, a tracked asset — that is the difference between converging and following.
A quiet channel
Messages are small and digitally signed, so the link stays cheap and stays trustworthy. Site data, drawings, photographs and map sets update themselves over the air rather than on a cable at the depot.
Handling can be finished from the tablet or from the desk. Whichever closes it, the other one knows.
The desk is not an island
Three places an event has to go that are not the operator’s screen.
Where an event goes nextThree destinations, and nothing below them.
The telephone
With the exchange connected, an incoming call opens the caller’s site before the operator has finished saying hello. Where the caller is responsible for several sites, those sites are listed to choose from. Outgoing calls are placed from the record itself, so nobody looks up a number while a client is waiting.
Third-party software
Events relay onward in Ademco ContactID over a persistent network connection or a paired virtual serial port, with a time correction where the two centres sit in different zones. Only events carrying a mapped code are sent — an unmapped one is not silently mistranslated — and an operator can forward one by hand when it is needed.
Panels nobody has registered
A panel transmitting against a site record that does not exist yet, or one entered with the wrong number, is kept in a separate archive rather than dropped. It is searchable by transmitted number, address and date, so a mis-commissioned installation is found instead of lost.
Afterwards
The evidence is a by-product of doing the work, not a report somebody writes on Monday.
- Every dispatch is in the log with its stages, the operator who handled it and the reason it was closed.
- The log filters by date and time, by stage, by operator and by reason; opening a row jumps to that event in the archive, in place.
- Closure reasons come from a list the centre maintains centrally, which is what makes a month of them countable rather than readable.
- The archive holds everything a site has ever sent, on the same record as the configuration that produced it.
What this is for
Attention where the consequence is
Priority follows what a signal means, not when it arrived, so the panic button never queues behind a missed test.
Fewer alarms that were never real
Commissioning, faults and floods are dealt with as configuration. The operator is left with the events that need a person.
Nothing reconstructed from memory
What was decided, by whom, and why it was closed the way it was, is written while it happens rather than remembered afterwards.
One account of the same event
Desk, vehicle and client are looking at one record in one state — which is what makes the conversation about it short.
Bring your own procedures.
Priorities, response instructions, closure reasons and suppression thresholds are yours to set, not ours to assume. Tell us how your centre runs a shift today and we will show you where it maps.
Talk to us