Friday, May 11, 2012

Time Delay Touch Switch

Fitting switches which turn off automatically after a preset time in these locations can reduce the electricity consumed quite significantly. A single 100W bulb left burning all night (say eight hours) would consume 800W/hrs. The same bulb fitted with a switch which turned off automatically after say five minutes would consume only 83W/hrs, even in the unlikely event that it were to be activated ten times during this period.

Triac Control
There are two major problems to solve when designing a touch switch to replace a conventional mechanical light switch and these concern the power supply and noise immunity.

The main switching element used to control the light is a triac, a semiconductor a.c. switch. This is normally in its off or non-conducting state but can be switched into conduction by means of a small trigger current fed to its gate terminal. Once triggered it will remain conducting until the current through it drops below a certain value (known as the holding current). This occurs normally on a.c. supplies when the voltage drops to zero at the end of every mains half-cycle, and the triac must be retriggered again if the light is to remain on.

As with any switch, the voltage across a triac when it is in its off state will be the supply voltage (in this case 230V a.c.). When the triac is on, the voltage across it will be around 1V a.c., depending on the current. The triac therefore dissipates very little power so no heatsink is required.
With normal house wiring, only the Live wire and its feed to the lamp are accessible at the wall-mounted light switch. The Neutral wire from the lamp is left inaccessible in the ceiling. Thus it is not possible to simply reduce and rectify the mains to provide a d.c. voltage for the triac control circuit, which must therefore be powered “through” the lamp, i.e. from the voltage across the triac.

An adequate d.c. voltage can be derived by delaying the point at which the triac is triggered in a mains half-cycle. This allows a higher voltage to develop across it which can be rectified and used to power the rest of the circuit.

The disadvantage is that it gives rise to EMI (electromagnetic interference) which can interfere with radio reception and computer systems. As a result, bulky and expensive chokes and capacitors have to be fitted to suppress it. This is unavoidable in light dimmers where we have to delay the triggering so that the power to the lamp can be controlled. For an on/off control, though, triggering should occur as early as possible so that full power is available to the lamp and without interference being generated.

In Control:
In this circuit, a small Zener diode is connected in series with the triac. This causes sufficient voltage drop so that after rectification a 3V to 4V d.c. supply is available for the control circuits. The Zener has the load (light) current flowing through it during operation and therefore dissipates power, but for the kind of currents used in lighting this only amounts to a few hundred milliwatts, a value well within the ratings of small wire-ended Zener diodes.

The circuit provides sufficient current to enable the triac to be triggered by d.c. pulses instead of the short a.c. trigger pulses normally used. As a result, the triac does not turn off between half-cycles, and so does not cause interference to the mains supply. This point is important as it also enables the circuit is to be used with low power fluorescent lamps if required.

Touch Control:
Touch control circuits, by their very nature, tend to be prone to spurious switching in response to the many glitches and pulses which normally occur on the mains supply due to other equipment being switched on or off. This is often difficult to overcome, but it can be very much reduced by building a delay into the circuit so that short pulses, which characterise most of the interference, are ignored and the circuit only responds to longer inputs.

By fitting a delay of a few hundred milliseconds, users will not notice any delay in switching, and interference, which rarely lasts for more than a few tens of milliseconds, will be largely ignored by the circuit. This is not a complete solution to the problem, but in this case there is also the advantage that even if the circuit should switch on accidentally, it will only remain on for a few minutes anyway!!!.

Circuit Description:
The circuit diagram for the Time Delay Touch Switch is shown in Fig.1. Triac CSR1 and Zener diode D4 are connected in series with the lamp as previously discussed. The voltage appearing across the Zener is rectified and smoothed by diode D3 and capacitor C4 to provide a d.c. supply of around 4V.
As this voltage only appears when the triac is on, capacitor C5 is connected as a capacitive mains dropper to provide a small current to the Zener to maintain the d.c. supply when the triac is off. Resistor R6 is included to limit the Zener current should the circuit be powered when the capacitor is discharged (very likely) and also to limit the discharge current which could destroy the triac when it switches on (also very likely).

The rest of the circuit is quite conventional and is based on a quad CMOS Schmitt trigger NAND gate, although the gates are used as logic inverters by having both inputs connected together. This type of gate has the advantage of switching cleanly once the input threshold voltage has been exceeded.

The input sensor is formed by placing a conducting foil behind the insulating front of a standard light switch blanking plate. The foil forms one plate of a capacitor. When it is approached by a hand, which becomes the other plate, the capacitor formed effectively connects the base of transistor TR1 to earth. (The human body is a fairly good conductor and exhibits a capacitance to earth of around 100pF, depending on size and area!)

Since the emitter circuit of npn transistor TR1 is connected via resistor R2 to the Live rail, which is at 230V a.c. with respect to earth, this is equivalent to a 230V input to the base of TR1 via the combined capacitance of the body plus that the hand and foil.

Not surprisingly, transistor TR1 turns on when the input polarity is correct and charges capacitor C1. Resistor R1 and preset resistor VR1 set the sensitivity of the circuit, allowing it to be activated by placing one finger or the whole hand on the plate. Resistor R2 discharges the capacitor when the hand is removed.

When the input threshold of gate IC1a is exceeded, the output switches low causing capacitor C2 to charge via resistor R3, which introduces the switch-on delay required to improve the noise immunity of the circuit. Should the output of IC1a go high again because TR1 turned on for only a short period due to noise, C2 would be quickly discharged via diode D1 so that the input of IC1b would not have time to go low.

Assuming a moderately long input period during which the output of IC1a remains low, however, IC1b’s input will also become low, causing its output to go high, so quickly charging capacitor C3 via diode D2. This, of course, causes the output of IC1c to go low and allow gate current to flow to the triac via pnp transistor TR2 and resistor R5. In turn, this causes the triac to be triggered and so switch on the lamp.

Note that the circuit uses negative triggering (i.e. a current flows out of the triac gate rather than into it) because triacs are generally more sensitive in this mode and require less gate current to ensure triggering.

When the hand is removed from the touch plate, diode D2 will become reverse biased and capacitor C3 will only be able to discharge via resistor R4 and preset VR2. These two components, together with C3, therefore determine the length of time for which the  triac remains triggered and the lamp turned on.

With the components specified, the turn-on time can be made variable from around 20 seconds to 2¼ minutes. This may be increased if required by increasing either R4 or C3, or both.

Note also that the time can easily be extended by touching the plate during the on period. This will recharge C3 so that if you stop on the stairs for a chat, you need not be plunged into darkness half way through the conversation!

In any event, even if this should happen, with the triac turned off 230V a.c. will appear across it and the small neon lamp LP1 will light. This is arranged to shine through a hole in the printed circuit board, illuminating the touch area so that it should be an easy matter to find the switch in the dark. Resistor R7 limits the current through the neon, which generally strikes at about 90V.

Touch Plate:
The touch plate assembly is made by using a piece of aluminium kitchen foil about 4cm square stuck behind the centre of an electrical blanking plate, as shown in Fig.2. Most blanking plates have a central area made of thinner plastic than the surrounding plate. This forms an ideal area for the neon light to shine through. (Do not make a hole in the blanking plate.) The aluminium foil should be removed from this area as shown in Fig.2. The surface of the foil should then be covered with strips of insulating tape.
A short piece of wire with about 1cm of insulation removed should be trapped beneath the insulation to enable the foil to be connected to the transistor on the circuit board. A small knot tied in the wire under the insulating tape may help to prevent the wire from being pulled out during final assembly.

Once this has been completed, the unit can be mounted in the wall box. The mains must be switched off before the existing switch is removed, and when making any adjustments. The two cables inside the wall box should be connected to the terminal block on the unit. Because of the way the circuit operates, it is important that the Live wire is connected to the L terminal of circuit.

This wire is normally coloured red but it may not be, depending on the installation. If you are not sure which of the two wires is the Live one, connect them either way around. If you have chosen the wrong one no damage will result but the circuit will not switch on and the procedure will need to be repeated with the wires reversed.

Note that modern installations will also have an earth wire which is not connected to the switch terminals but is usually connected to a terminal on the wall box. This should be left in place and not disturbed.

Once this is done, the switch can be secured to the wall box and the mains switched on. Make sure that the light can be switched on when the plate is touched and ensure that it goes out at the end of the previously set time.

If the circuit cannot be switched on, increase the sensitivity of the circuit by turning VR1 clockwise (switch off the mains before doing this). If it is still impossible to switch the circuit on, turn the preset fully anticlockwise and reverse the wires to the terminal block before repeating the procedure.

Do not increase the sensitivity beyond the point at which it switches on easily as this can make it more susceptible to random switching, or even prevent the unit from switching off at the end of the preset period. The latter can be increased by turning VR2 clockwise.

Should you wish to use the unit with a two-way switching circuit, to enable the lights to be switched on and off from two locations, this can be done quite easily with no re-wiring being involved. Fig.3 shows a typical two-way switching arrangement and the new switches may be used to replace the existing switches as shown.
The wire colours stated in Fig.3 are those normally used although they may be different depending on the installer. In any event, one of the existing wires will not be required and the touch switches should be wired in parallel as shown. It is important to ensure that the mains Live is connected to the correct terminal on each switch.

No comments:

Post a Comment