PIR sensor alarms get the picture
Paul Silcock, Technical Solutions Manager, Future Electronics (UK)
 |
READ THIS TO FIND OUT ABOUT:
 |
- Ideas for adding value to simple intruder alarm systems
- New image sensor, alarm enhancer and radio components to enable enhanced intruder detection designs
|
 |
|
|
Traditionally, the function of the Pyroelectric Infrared
(PIR) sensor in an intruder detection system (commonly
known as a burglar alarm) is simply to switch an output to
an alarm, which activates a siren and/or auto-dials a
police or security company control centre. Paul Silcock,
Technical Solutions Manager, Future Electronics (UK)
explains how to add functionality by adding image
capture without increasing size or the power budget.
Recent improvements in component technologies mean that it is
now possible to design a camera, flash and wireless transmitter into
the PIR sensor with only a small increase in the size of the end
product. By taking still or moving pictures of the intruder,
compressing them and then uploading the images to a central server,
the sensor greatly strengthens its deterrent effect against intruders by
providing evidence, which can be used against the perpetrator. This
adds a large amount of value to the end product, but without a
correspondingly large increase in the bill of materials.
The chief challenges involved in implementing such a design are,
first, to keep the physical size of the product to a minimum; and
second, to limit power consumption so that it can run for extended
periods on battery power. As this article will show, both of these
challenges can be met by using the latest generation of components.
The main principles of the design approach taken here are:
- To use highly integrated ICs, eliminating peripheral components and thus reducing footprint
- To keep large sections of the system in sleep mode except when
an intruder is detected. This means that the power-hungry
elements of the system, such as the short-range radio and image
processing, only operate at full power in short, infrequent bursts.

Fig. 1: Block diagram of enhanced PIR sensor system
The basic architecture of this enhanced intruder detection system is
simple (see Figure 1). When the PIR sensor detects a moving body in
its field of view, it wakes up the microcontroller.
The controller activates the camera flash and image sensor to capture
images of the intruder, and also activates a siren; the images are
processed, enhanced and compressed before being transmitted by
the short-range radio transceiver to a local control panel. The control
panel (not described in this article) will then dial out to a security
control centre, sending the images and an alert.
The first challenge for the designer is to find a way to implement
this considerable additional functionality, of capturing and wirelessly
sending images, without requiring a large amount of extra board
space. The answer is integration.
The output from a PIR sensor requires signal conditioning before
it can be processed by the controller (see Figure 2). Most intruder
detectors on the market today implement this signal conditioning
with a large number of discrete analogue components. Since the
controller functions required for this essentially dumb device are so
limited, such designs only need an extremely simple, very low-cost
8-bit microcontroller. No such microcontroller will offer the
sophisticated analogue functionality required to implement PIR
sensor signal conditioning. This is why these designs use a large
number of discrete analogue components.
Clearly, bolting on camera and radio functionality to such an
architecture would dramatically increase the size and cost of the end
product. There is an alternative, however: using a specialist mixed-signal
controller, the designer can absorb both signal-conditioning
and system-control functions into one device.
This will eliminate a large number of discrete
analogue components while providing the
capability to control the additional camera and
radio functions.
The effect of this integrated approach can be
seen clearly in Figure 3, an implementation
using the new ePIR controller from ZiLOG. The
ePIR solution uses a software algorithm for
signal conditioning, running on a
microcontroller from ZiLOG’s Z8 Encore! XP
family. In this intruder alarm design, the 8-pin
Z8F042ASB020SG device would be suitable. No
external gain or filtering circuitry is required,
which dramatically reduces component count.
At the same time, an ePIR design will offer
excellent immunity to false alarms and noise,
high reliability and integrated compensation
for environmental effects.
Because there is no filter, there is no loss of
signal, which results in improved range and
coverage. The total cost of an ePIR-based circuit
is cheaper than a traditional all-analogue PIR
sensor circuit, while offering better
performance than high-end digital PIR sensor
systems are able to provide.
An alternative approach to integrating the analogue signal-conditioning
circuitry is to use a member of the PSoC (Programmable
System-on-Chip) family of devices from Cypress Semiconductor. PSoC
devices are mixed-signal arrays of configurable digital and analogue
blocks, controlled by an on-chip 8-bit core. The blocks can be
configured to fulfil standard peripheral functions such as analogue-to-digital
converter, timer and UART. But the value of PSoC is its ability
to also integrate sophisticated analogue functions such as filters and
amplifiers.
The analogue functions of the conditioning circuit outlined in
Figure 2 can be integrated in the PSoC, meaning that a complete
solution comprising the PIR sensor, PSoC and a handful of tiny passive
components can be realised (see Figure 4). Furthermore, like the ePIR
implementation, the solution can be streamlined to fit into an 8-pin
device.
Such an implementation leaves processing bandwidth to spare to
control the downstream system functions: activating the image
sensor, power LED and image processor, and operating the short
range radio.

Fig. 2: Typical PIR sensor signal-conditioning circuit.
Enhancing and compressing images of an intruder
By integrating signal-conditioning functions in the motion-detection
section of the system, the board space that is released can now be
populated with devices to realise additional functionality. For instance,
capturing an image will require an image sensor, an image processor and
a small light-source.
The past few years have seen a proliferation of small, cheap image
sensors, thanks largely to the enormous volumes of such devices
consumed by the mobile phone industry. For this intruder detection
application, the MLX75007 camera-on-a-chip from Melexis would be
suitable. The device offers panoramic VGA resolution at 750x400
pixels, in black-and-white or colour. Originally developed for the
automotive industry, it is very competitively priced, while offering
high dynamic range and image quality that is good
enough to allow identification of an intruder, without being
of the high quality required by consumers in a camera
phone. The device has a Stand-by mode for reduced
power-consumption.
The images produced by this camera will result in the
creation of, typically, several hundred kilobytes of data. This
data cannot be permanently stored locally, otherwise the
intruder could simply remove the sensor to get rid of the
evidence of the crime. So the images have to be
transmitted to a safe location. Wireless transmission is
preferred, as it reduces installation costs and eliminates any
risk of a cable being cut by an intruder. But transmitting
large data files wirelessly consumes a lot of power and
takes a lot of time. The answer to this problem is to use an
image-processor to compress the data before transmission.
As shown above, when designing the motion-sensing
function, integration allowed a big reduction in
component-count.
The same principle can be applied to the image-processor.
A useful device here is the tiny, 8mm2,
BU6569GVW from Rohm: it offers interfaces for a
2 Megapixel camera, NAND Flash and SD/MMC cards.
Crucially it also provides adaptive differential pulse-code
modulation and JPEG codecs for image compression.
It is worth bearing in mind the requirements of this
application: the subject of the image does not want to
be photographed and could be moving, the pictures
could need to be taken at night, and yet to be useful
they must allow the subject to be clearly identified.
The design therefore requires two additional functions:
flash lighting and image enhancement. A power LED
provides a bright burst of light from a low-voltage source
while occupying a small area. The LXCL-PWF3 Flash LED
from Philips Lumileds measures just 2mm x 1.6mm. A single device can
illuminate up to a distance of 2m with a 1A pulse.
Unmodified, the signal from an image sensor under the harsh lighting
conditions experienced by typical security applications will often be
unusable. Areas of excessive brightness or darkness will cause the image
sensor to bleach or black-out the image.
Typical image enhancers work on the principle of adjusting the balance
of the whole image. For this application, however, a better choice would
be the BU1570KN adaptive image enhancer from Rohm. This device uses
hardware-based image-processing technology to raise or lower the
brightness of only selected portions of the image which are too bright or
too dark. This approach achieves much better image clarity under
extreme lighting conditions. The BU1570KN also offers motion detection,
colour correction and brightness determination. In a PIR sensor
application, the motion detection hardware can verify inputs from the PIR
sensor.

Fig. 3: Board space saving achieved by use of a highly-integrated ePIR controller from ZiLOG to replace a
traditional analogue PIR sensor design.
Transmitting images for secure storage
The enhanced, compressed images now need to be sent to a central
server for secure storage, where they cannot be tampered with by an
intruder. The main design challenge is, again, to provide this functionality
with a miniature, power-efficient device.
The data rate requirement is quite moderate. JPEG typically achieves a
10:1 compression ratio with little perceivable loss in image quality. This
means that a typical 500kB image file could be compressed to 50kB, and
transmitted in 2s at 250kbit/s, even with a prudent allowance for
overhead.
A sensible approach would be to buffer images in a small Flash
memory on activation of the sensor, bringing the NAND Flash interface in
the BU6569GVW into play, and then stream them to the server.
Operating in Europe’s unlicensed 433MHz and 868MHz bands, Micrel’s
MICRF505/6 transceiver comes in a 5mm2 MLF package. It provides a
data-rate up to 200kbaud and draws 28mA in transmit mode.
The use of just a transmitter would give a small saving in board space
and power, whilst a transceiver would allow for confirmation of receipt
of transmitted images. Bearing in mind the application, the risk of failed
transmission is not worth taking: the system might only ever be called
on to operate once – when it does, the user will expect it to work
flawlessly.
For end products that are to be marketed worldwide, the 2.4GHz
frequency is more suitable. Here, Freescale’s MC13202 is compatible with
the IEEE 802.15.4 standard and has an over-the-air data rate of 250kbps
using Direct Sequence Spread Spectrum (DSSS) coding for interference
immunity.

Fig. 4: PIR sensor signal-conditioning circuit implemented in a CY8C27443 PSoC.
Optimising this design for low-power operation
It is inevitable that, as devices are added to a design in order to add
functionality, so more power is consumed. But the enhanced
marketability of the improved PIR design would be negated if an
expensive and bulky power supply were also required.
Fortunately, however, recent advances in DC-DC conversion
technology come to the design’s rescue. New, highly efficient
synchronous buck regulators with integrated FETs, and in some cases
inductors, provide considerable savings in battery-powered applications.
For instance, the SupIRBuck family of Point-of-Load (PoL) voltage
regulators from International Rectifier integrates high performance
synchronous buck control ICs and HEXFET trench-technology MOSFETs
in a compact 5mm x 6mm Power QFN package. The IR3812MPbF, for
example, can supply up to 4A from a wide-input range of 2.5V to 21V,
with efficiency exceeding 90%.
Fairchild Semiconductor’s TinyBuck™ family of synchronous buck
regulators would also be very suitable for PIR sensor applications. The
FAN2103, for example, supplies up to 3A from a 3V-24V input in a 5mm x
6mm 25-pin MLP package. Efficiency can be in excess of 95%.
Improvements in the efficiency of power ICs are not the only way for
this design to hit its power budget. In this application, most of the
system is active very infrequently, and when it is active it is in a short
burst. The key to long battery life, therefore, is to use the controller to
put most of the system into sleep mode while waiting to be activated
by the PIR sensor.
On activation of the PIR sensor, the microcontroller will supply power
to the camera, LED flash light, image processor and Flash memory to
enable capture, compression and storage of images. With the exception
of the Flash memory, these devices will all be switched off before the
radio transceiver is powered up and the streaming of image data can be
initiated. Finally, the system will return to its inert state awaiting the next
activation.
The average current drawn by
this system, of an estimated 6mA
as opposed to around 0.25mA for
a traditional dumb PIR sensor, is
moderate, and so ensuring long
battery-life should not be
technically challenging. In fact,
the more important calculation is
whether the chosen battery can
support the peak load, when the
system is fully active. To be sure
of this, the design engineer must
carefully work out current
consumption at each stage of
operation.

Table 1: Estimated bill of materials comparison between a traditional
dumb PIR sensor and the enhanced camera/sensor with radio.
Conclusion
Improvements in a range of
components recently brought to
market have, in combination,
enabled a valuable enhancement
to traditional dumb PIR sensors
without sacrificing battery life
and without increasing significantly the size of the end product. For
designers prepared to consider non-traditional controllers such as the
ePIR from ZiLOG or the PSoC from Cypress, such a design is relatively
easy to implement.
Table 1 compares the component costs of the traditional solution
and the enhanced system. While the cost of materials is higher in the
improved design, the extra value it provides, derived from the ability
not just to detect but also to identify an intruder, should support a
much higher selling-price.
©Copyright 2008 Future Electronics Ltd. All trademarks contained here in are the property of their
respective owners. All applications for samples, badge boards, demo kits and other
advertised materials from Future Electronics are subject to qualification.