# Simple Alarm System

Schematic Diagrams      Comments Off on Simple Alarm System

The circuit presented
here is a very simple and yet highly effective alarm system for
protecting an object. The circuit requires no special devices and can be
built using components that you will no doubt be able to find in the
junk box. The alarm-triggering element is a simple reed switch. To
generate the alarm signal itself any optical or acoustic device that
operates on 12 V can be used: for example a revolving light, a siren, or
even both. In the quiescent state the reed switch is closed. As soon as
the reed switch opens, the input to IC1.B will go low (previously the
potential divider formed by R2 and R3 held the input at 5.17 V, a logic
high level).

A turn-on delay of between 0 and approximately 90 s can be set using
P1, and a turn-off delay of between 0 and approximately 20 s can be set
using P2. When the system is turned on (using S1), the turn-on delay is
activated, giving the user of the system at most 90 s to leave the
object alone before the system goes into the armed state, and the object
is then protected. Once the reed switch opens the turn-off delay of at
most 20 s starts: this allows the rightful owner of the object to turn
the system off before the alarm is triggered. If some unauthorised
person causes the reed switch to open, the alarm will be triggered after
the turn-off delay.

Also, even if the reed switch is briefly opened and then closed
again, the alarm will still be triggered. Once the alarm is triggered,
T3 will conduct for about 45 s (because of R8 and C5). The turning off
of the alarm is necessary to avoid the nuisance caused by a permanently
sounding alarm system. The system then returns to the armed state, which
means that the next time the reed switch is opened the alarm will
trigger again. If it is not desired that the duration of the alarm be
limited, for example if a visual indication is used, D5 should not be
ﬁtted. The system can be extended by ﬁtting multiple reed switches in
series. As soon as any one is opened, the alarm is triggered.

When S1 is closed C3 charges via P1. Depending on the potentiometer
setting, it takes between 0 and 90 s to reach the input threshold
voltage of IC1.A. The output of IC1.A then goes low and D3 stops
conducting. Assuming the reed switch is closed, the inputs of IC1.B stay
high and the output therefore low. If the reed switch is opened after
the turn-on delay expires the output of the gate will change state and
turn on T1. This ensures that the output of the gate remains high even
after the reed switch is closed again. C4 now starts charging via P2,
reaching the input threshold voltage of IC1.C after between 0 and 20 s,
again according to the potentiometer setting. The output of IC1.C goes
low, and T2 and T3 are turned on – and the siren sounds.

Any Darlington transistor can be used for T3. At the same time, C5
charges via R8, reaching the input threshold of IC1.D in about 45 s.
When the output of IC1.D swings low, it pulls the inputs of IC1.A low
via diode D5: the siren stops and the system returns to the armed state.
If the potentiometers P1 and P2 are replaced by ﬁxed resistors it is
possible to build the circuit small enough to ﬁt in a match-box, without
the need to resort to SMD components. This is
ideal if the circuit is to be used to protect a motorbike. If the alarm
system is to be used in a car, an existing door switch contact can be
used instead of the reed switch. In this case an RC combination needs to
be added to prevent false triggering. Use a 10 µF/25 V electrolytic for
C6, a 100 kΩ resistor for R9 and a 1N4001 for D7. It is again possible
to wire multiple door switch contacts in parallel: as soon as one
contact closes, IC1.B will be triggered.