Burglar Alarms and Power Failure

One of the problems mentioned in Y2K planning is that of burglar alarms. It is worth looking into this, not least because many of us have alarms protecting our homes. There are two possible causes of problems: computers in the alarms and mains failure. Note that this article describes what I think to be the case - I have not contacted alarm manufacturers to confirm my reasoning.

Many modern alarms have embedded processors, i.e. small computers contained within them. Probably few domestic alarms have any awareness of the date; certainly mine has no way of setting it. If they live in blissful ignorance of the date, it seems unlikely that they will suffer from direct Y2K problems in the sense of program bugs.

Some alarms are more complicated. Many bank vaults, for instance, have time locks which will only open during scheduled opening days and hours, to prevent them being compromised by robbers at night or during weekends. These might certainly have Y2K problems, so it is probably worth taking a lot of money out before Christmas 1999 . Some shops and other commercial premises might have date-dependant alarms, but these are unlikely to be our problem (unless they sell essentials like Chorley Cakes).

For home alarms at least, probably the main danger is that of mains failure. The electricity supply utilities are putting a lot of effort in preparing for the millenium problem, but power failures are still a possibility. How this affects domestic burglar alarms depends on their design and the duration of the power outage.

Part of the circuit of a typical alarm is shown in the figure. Again, I have not made any survey of alarms in general, but I suspect this to be the general pattern. The diagram really just shows the power supply arrangements, somewhat simplified. The thing to note is that there are three parts to the power supply.

The primary supply is the mains; normally this is present. If it fails, the 12V battery in the control panel takes over. This has two benefits - a burglar cannot disable the alarm simply by cutting the mains and the system can survive accidental power failures. The 12V battery is float-charged. Note that a mains failure is not treated as an indication of intrusion - it is assumed that adequate intrusion sensors have been installed.

The siren unit (few alarms use an actual bell nowadays) is mounted remotely from the control panel and has its own backup battery which is float-charged from the 12V line. This is not intended to guard against accidental power failures. It is there so that the alarm cannot be disabled merely by cutting the cable from the Control Panel to the Siren. The control lines between the two are arranged so that any such simple cutting will set off the siren. (Incidentally, I say 'simple cutting' because the wire colour coding tends to be standardised and it is quite easy to disable these systems by cutting back the insulation and shorting the right wires. Any burglar who examines their own alarm would know what these are. If you install such a system, make sure that the wires to the bell box aren't accessible from outside the house.)

Now let us consider what happens if the mains fails. While the 12V battery has enough voltage, everything is fine. If the mains remains off, this supply will drop below 12V. It may have to drop quite a way as circuits may work from lower voltages derived from it. (Mine runs the embedded processor from a 5V regulator driven from the 12V battery, so this could lose half its voltage before the processor mis-behaved.) The 12V line also feeds the siren unit, however. When this supply falls below a certain voltage, the siren will decide that the system has been tampered with and will go off. Probably all such sirens have a time-out: they will not sound until their battery fails. Mine (which is probably typical) has a choice of 2 minutes or 15 minutes; after this the timing circuit will stop the noise, much to the relief of the neighbours. (I set mine to 2 minutes, on the grounds that if a burglar isn't scared away by then they're not going to be.) If the siren's battery is in bad condition it might fail earlier, of course.

Now let us look at some specific figures for actual alarms. The 12V battery seems typically to be a 1.2Ah for domestic premises, though a 6Ah appears common for commercial ones. These are normally sealed lead-acid units,Yuasa or similar. My alarm consumes about 150mA from the 12V line normally. The amount will depend on the installation, e.g. the number of PIR (passive infra-red) sensors in use. The consumption goes up to about 530mA when the alarm is sounding. These figures were measured by turning off the mains, pulling a wire off the battery and putting an ammeter between that wire and its battery terminal.

From these figures we can calculate how long the system will survive without mains; the results are given in Table 1. Columns are given for a range of 12V amp-hour capacities, described as the two battery types with all, half and quarter their original capacity remaining (i.e. allowing for old batteries). The 'No Alarm' row shows how long batteries will last if the alarm does not go off. If it does go off, then 2 minutes will require an extra 18mAh and 15 minutes 133mAh. The bottom two rows show the battery times reduced to allow these siren durations. These figures are specific to my installation; they should be regarded merely as giving a general idea of actual figures. If the actual times matter to you, you should repeat the measurements on your own alarm. Note that when you break the power the alarm will sound. It is well to revise the instructions for silencing the thing before you start! You will also lose all the settings and have to reset the user codes, entry/exit zones, etc., afterwards.

You will note that the backup battery in the siren has been ignored. If it were in good enough condition, it could be relied upon to provide all the siren current. The bottom two rows could then be replaced by the 'No Alarm' row. My siren battery has no identification marks, but it looks horribly like a 6V, 280mAh NiCd. Nickel-Cadmium batteries are about the worst type to use in float-charging situations. Mine has been there for about five years; I would expect it to have about as much capacity as a slightly large electrolytic. If I relied on it I would replace it by something more useful such as three 2V Cyclon cells.

In fact, the benefit of the siren's battery to the times above is negligible except for the two left-most columns of the bottom row. These columns refer to 12V batteries which have reached the ends of their useful lives, so the sensible reaction would be to replace the 12V battery, not use the figures.

My analysis of these figures is that 1.2Ah batteries are quite useless for the sort of outages that Y2K might cause. They are fine for the majority of power failures nowadays, which last anything from a second to half an hour or so. If things go badly, we could easily be looking at a day or two without electricity. The problem is that the battery might go flat after you had been called out for a long shift. I think there are three obvious options if we do have lengthy power cuts.

1. The alarm could be de-commissioned for the duration. This might not be stupid - after all, alarms only warn of intruders, they do not prevent them. With the likely number of false alarms from other houses your alarm would probably be ignored.
2. The 1.2Ah battery could be replaced with something larger. One of the 6Ah packs, for instance, would do about a day and a half. It might be too large to fit inside the control unit. Autumn 1999 would seem a good time to do this, so the battery would still be in prime condition at the year turn.
3. The alarm could be powered from whatever 12V source you are using for RAYNET operations during power failure. This would require electrical compatibility (can you earth the -ve side of the alarm?) It would also risk the alarm going off when your generator runs out of petrol at four in the morning!

In summary, domestic burglar alarms are unlikely to suffer directly from software problems brought on by the millenium. They may be vulnerable to extended power failures, however. My very limited tests suggest that typical units will survive 8 hours at most without upgrading. Anyone who has such an alarm would be advised to examine it and make plans for power failures, both to protect their property and to avoid annoying neighbours.

Chris Trayner G4OKW

G4RDL