Return to site index

A few suggestions before you spend money on a CCTV system

No CCTV system is cheap, so do not waste money on the wrong system. It should give high quality recordings suitable for identification and use in court, not just a blurred image, together with the correct date and time.

The ideal image will be well lit, as detailed as possible, from a camera at around head height, and not distorted. Unfortunately the camera could be vandalised, so the optimum may be just out of reach, at a minimum height of around 2.5 metres. The best lighting conditions may only occur at a particular time of day, so consider two cameras looking from opposite directions.

Modern cameras may be extremely sensitive, but do not expect the same quality under low light conditions. Reading vehicle number plates will be difficult if it is moving or the vehicle lights are on, so an LED floodlight with passive infra-red PIR detector will help. Some recorders can provide a switched alarm output that could be used to drive a relay.

The two most common shapes, dome and bullet, will probably have identical contents. Most CCTV cameras have a fixed wide-angle lens, giving wide coverage, all in focus, but a usable image only when relatively close. Different lenses may be available, but more expensive and perhaps with smaller apertures. A lens with longer focal length will have a more restricted depth of focus. External housings are available with an integral heater to prevent misting and condensation problems, and may be useful with narrow angle lenses.

Many cameras have infra-red LEDs for night vision, but they are often very close to the lens. The LEDs attract insects, and the insects attract spiders that will spin a web right in front of the lens. You may need to spray the camera with SpiderX. The bullet shape may have a lens hood, but possibly too small to do much without being visible. I have found that a dome camera under an inverted length of Ogee guttering stays cleaner for longer, but although out of view the cover could reflect light from the LEDs and cause fogging under low light conditions, cured using less reflective adhesive tape or a flat cover. I have now cut some from surplus 204mm wide ventilation duct.

Some cameras can handle a very wide dynamic range WDR of light levels, and can give an acceptable full colour image at very low light levels. Good quality wide aperture lenses can be expensive. There may be a choice of lenses, including remote controlled optical zoom, although some are designed to be manually adjusted and re-focused only on installation. Some also provide automatic or remote control of pan, tilt, and zoom, audio facilities, and may have facial and number plate recognition systems.

Light intensity drops rapidly away from a light source, so consider multiple lights around the area to be covered.

Modern high definition cameras give one or more compressed video streams over ethernet cable, not co-ax, and can run on either separate local power (normally 12 volts DC at the camera), or Power over Ethernet, PoE, which sends a nominal 48 volts DC through the cable.
Many cameras are designed to connect via 10/100 ethernet cable using just two pairs, while the other two pairs accept PoE power.
Wireless cameras still require power, although may be able to run on solar power with battery backup.

The digital compression software can easily perform motion detection at the camera, with sensitivity control and masking to disable selected areas. This can work well, although it may be difficult to ignore reflections and shadows from anything moving outside the enabled areas. About half of my cameras are close to busy roads, with little difference between motion detection and 24/7 continuous recording. The data stream can be delayed in memory for a few seconds, allowing the motion detection signal to start the recording from before the motion was detected, continuing for a timed period after detection ends.

Cameras often have software that can send their output via the internet if required, and some recorders are happy to receive all the camera outputs via a single network cable, so that the cameras can be connected to, and PoE powered by, a network switch in another more convenient location such as a house loft. If a recorder supplies direct PoE to a limited number of cameras then the maximum number of cameras that will be accepted via the network is likely to be the specified total less the number of PoE connectors provided. Recorders may claim to work with cameras manufactured by others, but perhaps not under all connection modes, so possibly restricted to network connection only. Remote control software, and remote access to the recorder, are normally provided, but its use may be restricted to specific operating systems. Be aware that almost anything connected to the internet could be a security risk, and possibly used by others to obtain information such as whether premises are likely to be occupied at particular times.

There is usually a choice between high definition and high frame rate, so specifications contain confusing "up to" information. The number of pixels built into the Charge Coupled Device CCD optical sensor set the maximum possible definition, but the processing, compression, and transmission software is able to limit the transmission data rate as required.

The recorder will also be subject to various limitations, including the maximum number of pixels per frame and the frame rate for some cameras, the maximum total numbers for all cameras, and the maximum network data rate, so may also be specified as "up to" a maximum number of cameras, but effectively limited to a lower number if best quality and frame rate are required.

Any high definition system will provide a huge amount of data, and recorders can be subject to overheating and early failure, so choose a cooler location, and shield them from direct radiation if located in a loft.

Many modern CCTV manufacturers produce cameras and recorders that will provide additional features when all are their own design and manufacture, but will also be able to use an interoperability system such as ONVIF or PSIA. There are various ONVIF specifications for a wide range of products, so more information is required than just ONVIF on the box to know what will work, including the ONVIF specification date. There is more information on http://www.onvif.org.
Many now use specially designed low power chips with ARM processors. I have found that although a recorder may specify a maximum disk capacity, it may work perfectly with a much larger disk that has already been partitioned using GPT and then formatted using the Linux ext4 format, although you may not wish to chance the cost of a very high capacity disk designed to handle 24/7 CCTV.

The recording medium has to record the cameras as they are presented. Special surveillance hard drives can cope with heavy loading 24/7, while solid state drives must be able to cope with repeated re-writing. The amount of storage space required will vary according to location, and can be reduced if recording is only started after motion is detected. My original 2TB drive would only retain about 3 to 5 days recording before it was over-written, while 1TB might retain recordings for a month in a quiet area where the only movement is due to wild animals and birds, with few rapid lighting changes and little wind blown vegetation. Some cameras can also record to an internal memory, and MicroSD cards specially designed to record high data rates 24/7, with a stated minimum number of repeated over-writes, are available.
Current Solid State Discs (SSD's) have an expected limit to the number of times they will be able to accept a write, which may be OK for general use but can cause early failure when used to record CCTV images.

Some companies purchase off-the-shelf hardware and install their own software, possibly built using sample software from the original manufacturer, and then sold under their own brand name. I have found that while software update procedures may be provided by some manufacturers, customer service and actual updates may be unavailable even to their main distributors.

There are many different types of ethernet cable. CCA means copper coated aluminium which is cheap, but will quickly corrode and fail. The standard cable is designed for internal use only, and should be protected if used outside. External cable is available, but more expensive, as is low smoke, low halogen cable which may be required under local building regulations. Cat5e cable will probably be adequate, although Cat6 cable has thicker conductors and will be suitable for longer lengths and higher future data rates. A fixed installation will not require the more expensive stranded rather than solid conductors, but ensure that the connectors are designed to match the cable. Insulation displacement connectors only for use with stranded cable may have two offset spikes, but three spikes with alternate offsets are more reliable with solid cored cable.
Ethernet cable is notorious for kinking and snagging, so is normally supplied in a box designed to supply the cable from the centre of the spool. Do not pull the cable from the box until you are ready to pull it straight into position. Make sure that all connections are protected from damp as the contacts can be dissolved if wet while powered. Totally waterproof silicone oil or grease can help protect the cable ends and connectors.

The correct standard cable connections are important but not immediately obvious. A simple tester may show that the correct pins are connected end to end, but they must also have the cable pairs connected via the correct pins.

Standard ethernet connections (pdf file)

Most camera CCD chips have each pixel divided into areas sensitive to just one colour, with some areas totally insensitive to light. Broadcast quality cameras have a beam splitting prism known as an Ice block behind the lens, with multiple thin surface layers able to pass or reflect individual colours, similar to the effect seen when oil is floating on water. CCD chips are then bonded in place to sense only the individual colours. Each pixel on a chip may still have some insensitive areas, so microscopic lenses are included to focus all the available light on to only the sensitive areas. The entire uncompressed output is then recorded or used live to minimize patterning during later processing and mixing. Final compression to normal broadcast standards is applied at the transmitter.

Return to site index