Multiway switching
Multiway switching is a method of connecting switches in groups so that any switch can be used to connect or disconnect the load. This is most commonly done with lighting.
Two locations
Switching a load on or off from two locations (for instance, turning a light on or off from either end of a flight of stairs) requires two SPDT switches. There are two basic methods of wiring to achieve this, and another not recommended.
In the first method, mains is fed into the common terminal of one of the switches; the switches are then connected through the L1 and L2 terminals (swapping the L1 and L2 terminals will just make the switches work the other way round), and finally a feed to the light is taken from the common of the second switch. A connects to B or C, D connects to B or C; the light is on if A connects to D, i.e. if A and D both connect to B or both connect to C.
The second method is to join the three terminals of one switch to the corresponding terminals on the other switch and take the incoming supply and the wire out to the light to the L1 and L2 terminals. Through one switch A connects to B or C, through the other also to B or C; the light is on if B connects to C, i.e. if A connects to B with one switch and to C with the other.
If the mains and the load are connected to the system of switches at one of them, then in both methods we need three wires between the two switches. In the first method one of the three wires just has to pass through the switch, which tends to be less convenient than being connected. When multiple wires come to a terminal they can often all be put directly in the terminal. When wires need to be joined without going to a terminal a crimped joint, piece of terminal block, wirenut or similar device must be used and the bulk of this may require use of a deeper backbox.
Using the first method, there are four possible combinations of switch positions: two with the light on and two with the light off. N.B.
An unrecommended method
If there is a hot (a unique phase) and a neutral wire in both switches and just one wire between them where the light is connected (as in the picture), you can then solve the two way switch problem easily: just plug the hot in the top from switch, the neutral in the bottom from switch and the wire that goes to the light in the middle from the switch. This in both switches. Now you have a fully functional two way switch.
This works like the first method above: there are four possibilities and just in two of them there is a hot and a neutral connected in the poles of the light. In the other ones, both poles are neutral or hot and then no current flows because the potential difference is zero.
The advantage of this method is that it uses just one wire to the light, having a hot and neutral in both switches..
The reason why this is not recommended is that the light socket pins may still be hot even with the light off, which poses a risk when changing a bulb. Another problem with this method is that in both switches there will be hot and neutral wires entering a single switch, which can lead to a short circuit in the event of switch failure, unlike the other methods.
This method is in defiance of the National Electrical Code (USA) and the Canadian Electrical Code. In nearly any and all applications, neutral conductors should never be switched. Not only is this a shock hazard due to mistakenly believing that a hot conductor is switched off; it is also a fire hazard and can destroy sensitive equipment due to excessive and unbalanced current flowing on hot conductors that would otherwise flow back to ground on the neutral conductor.
More than two locations
For more than two locations, the two cores connecting the L1 and L2 of the switches must be passed through an intermediate switch (as explained above) wired to swap them over. Any number of intermediate switches can be inserted, allowing for any number of locations.
As mentioned above, the above circuit can be extended by using multiple 4-way switches between the 3-way switches to extend switching ability to any number of locations.
Power switching
When a switch is designed to switch significant power, the transitional state of the switch as well as the ability to stand continuous operating currents must be considered. When a switch is on its resistance is near zero and very little power is dropped in the contacts; when a switch is in the off state its resistance is extremely high and even less power is dropped in the contacts. However when the switch is flicked the resistance must pass through a state where briefly a quarter (or worse if the load is not purely resistive) of the load's rated power is dropped in the switch.
For this reason, most power switches (most light switches and almost all larger switches) have spring mechanisms in them to make sure the transition between on and off is as short as possible regardless of the speed at which the user moves the rocker.
Power switches usually come in two types. A momentary on-off switch (such as on a laser pointer) usually takes the form of a button and only closes the circuit when the button is depressed. A regular on-off switch (such as on a flashlight) has a constant on-off feature. Dual-action switches incorporate both of these features.
Inductive loads
When a strongly inductive load such as an electric motor is switched on input surge current which may be several times larger than the steady current flows. When switched off, the current cannot drop instantaneously to zero; a spark will jump across the opening contacts. Switches for inductive loads must be rated to handle these cases. The spark will cause electromagnetic interference if not suppressed; a snubber network of a resistor and capacitor in series will quell the spark. Exact values can be optimised for the particular application, but for many cases a 100 ohm resistor in series with a 100 nanofarad capacitor will do.
Up or down
Domestic light switches are generally moved up to switch on in the USA, but down in most of Europe and Australia.
The reason for the difference remains a bit of a mystery. A few hypotheses are often put forward, (for example in the USA if the switch spring fails it cannot cause the switch to accidentally turn on, in other words it will fail safe), but none have been validated. Since there is no significant technical reason for either preference, the standards likely developed due to chance and some degree of cultural isolation.
In countries prone to earthquakes, such as Japan, most switches rock sideways to prevent the switch from inadvertently being turned on or off by falling objects.
Contact bounce
Contact bounce (also called chatter) is a common problem with mechanical switches and relays. Switch and relay contacts are usually made of springy metals that are forced into contact by an actuator. When the contacts strike together, their momentum and elasticity act together to cause bounce. The result is a rapidly pulsed electrical current instead of a clean transition from zero to full current. The waveform is then further modified by the parasitic inductances and capacitances in the switch and wiring, resulting in a series of damped sinusoidal oscillations. This effect is usually unnoticeable in AC mains circuits, where the bounce happens too quickly to affect most equipment, but causes problems in some analogue and logic circuits that respond fast enough to misinterpret the on-off pulses as a data stream.
Sequential digital logic circuits are particularly vulnerable to contact bounce. The voltage waveform produced by switch bounce usually violates the amplitude and timing specifications of the logic circuit. The result is that the circuit may fail, due to problems such as metastability, race conditions, runt pulses and glitches.
There are a number of techniques for debouncing (mitigating the effects of switch bounce). They can be split into wet contacts, timing based techniques and Hysteresis based techniques.
Wet contacts
Mercury wetted switch contacts do not suffer from bounce, as once the connection is made the mercury keeps the contact conducting during mechanical bounce.
Mercury wetted switches are not a popular option today, primarily due to mercury's toxicity.
Timing based
Resistor and capacitor
If an on/off switch is used with a pull up (or pull down) resistor and a single capacitor is placed over the switch (or across the resistor, but this can cause nasty spikes of current on the power supply lines) then when the switch is closed (generally pressed) the capacitor will almost instantly discharge through the switch. But when the switch is opened (generally released) the capacitor takes some time to recharge. Therefore contact bounce will have negligible effect on the output. The slow edges can be cleaned up with a Schmitt trigger if necessary. This method has the advantage of fast response to the initial press but the current surges through the switch may be undesirable. Other RC based systems are also possible with various responses and such systems are probably the easiest method when constructing with simple logic gates and discrete components.
State machines and software
A finite state machine or software running on a CPU can be designed to wait a fixed number of clock cycles after any transition before registering another one. This provides a cheap option for debouncing when a microprocessor, microcontroller or gate array is already in use but is unlikely to be worthwhile if constructing with single logic gates CLPD's.
Sampling
Arguably the simplest way to debounce a switch transition, either in hardware or software, is merely to sample the switch state at intervals longer than any possible train of bounces. This guarantees that any bouncing affects at most one sample, which must agree either with the previous sample or with the following sample. Either case results in only one clean transition in the sampled data. A simple hardware implementation is a single D-type flip-flop clocked at a suitable rate; and software sampling is easy to program. For most switches, a suitable sampling rate can easily introduce less latency than a human being can perceive.
Hysteresis
Alternatively, it is possible to build in hysteresis by making the position where a press is detected separate from that where a release is detected. As long as the bounces are small enough not to take the switch between these positions, bounce problems will be eliminated. Hysteresis can be mechanical or electronic (e.g. a Schmitt trigger).
Changeover switch
A changeover switch provides two distinct events, the making of one contact and the breaking of the other. These can be used to feed the inputs of a flip-flop. This way the press will only be detected when the pressed contact is made and the release will only be detected when the released contact is made. When the switch is bouncing around in the middle no change is detected. To get a single logic signal from such a setup a simple SR latch can be used.
Variable resistance
Normal switches are designed to give a hard on-off but it is also possible to design one that varies more gradually between the hard-on and hard-off states. This keeps the output changes caused by bouncing small. Then by feeding the output to a Schmitt trigger the effect of those bounce based changes can be eliminated.
NOTE:
commercial aircraft components
