Before an electrical circuit interruption process is initiated – that is, when the contacts of an interrupting device start to open or the injection of mobile carriers into a semiconductor switch is restricted – the interrupting device must first make a trip/no-trip decision.  The period of time between the initiation of an overcurrent condition within a circuit and the initiation of interruptive action by the circuit protection device is termed the detection period.  The different types of protection devices detect overcurrents in different ways.  Thus, they can have different detection periods for the same overcurrent conditions. 

The detection mechanism in a fuse is the melting and the vaporization of a fusible link.  In a thermal breaker, dissimilar metals, bonded together along a single surface, expand differently under the direct or indirect resistive heating of the overcurrent.  This forces a lateral mechanical movement, perpendicular to the bonded surface, which releases a latched contact separation mechanism.  In some types of thermal breakers, the contact mechanism can be formed using the bi-metal material itself.  In these devices, the bi-metal arms/contacts snap open when they absorb sufficient energy from the circuit overcurrent.  Another form of thermal breaker utilizes the longitudinal expansion of a hot wire, which carries the overcurrent, to release a contact latch.

The detection portion of a magnetic breaker is comprised of an electromagnet driven by the circuit current.  An overcurrent will develop, within the electromagnet, enough magnetic pull to trip a spring restrained latch which, as in the thermal breaker, allows the spring loaded contacts to separate.

A solid-state switch detects overcurrents electronically, in many cases by simply monitoring the voltage drop across a low value resistance which carries the circuit current. 

Obviously, the faster a protection device can detect an overcurrent, the shorter the detection period.  But, in the majority of cases, the fastest possible detection speed is not desirable.  The speed of detection must be controllable and inversely matched to the severity of the overcurrent.

Series-connected protection devices must be coordinated.  For a given level of overcurrent, the device nearest to, and upstream from, the cause of the overcurrent must have the fastest response.  Devices which are further upstream must have a delayed response, such that the minimum circuit removal principle is adhered to.  When we speak of response, we are referring to the total response time, or total clearing time, of the interruption device, from the time of the overcurrent initiation to the final current-zero at which interruption is completed.  Since it is far easier to engineer the extent of the detection period for a given level of overcurrent than it is to control the extent of the actual current interruption process, the total response time of any protection device is, by design, determined principally by the size of, and the time required to detect, the overcurrent state.

The interruption period is defined as the length of time between the start of interruptive action – for example, when the contacts start to part – and the final current-zero.  The sum of the detection period and the interruption period is then the total clearing time, or total trip time, of the protection device.  These different time periods are shown in figure 1.5.

In contrast to the detection period, the interruption period cannot be engineered to decrease the intensity of an overcurrent increases.  The interruption period is, however, almost always designed to be as short as possible, since during this period the protection device is absorbing energy, due to the overcurrent flowing through the voltage drop across the contacts (or terminals in the case of a solid-state device).  If protection devices, other than fuses, do not clear the overcurrents fast enough during this period, they can be destroyed due to their own power dissipation.  Of course, fuses by design are always destroyed when they interrupt a circuit.

In AC circuits, the interruption period will last to either the first forced current-zero or the first natural current-zero at which the switching medium (arc or solid-state material) can reach its non-conducting blocking state.  In DC circuits, the current-zero state is always a result of a forcing action by the interrupting device. 
There are additional time periods of interest during the current interruption process, such as contact travel time, arc restrike voltage transient time, thermal recovery time, and charge storage time (for solid state devices). 

The voltage and current in a complete electrical circuit obey Kirchhof’s voltage and current laws.  These laws simply stated are:  the rises and drops in voltage around any closed circuit (a circuit loop) must sum to zero; and the total current flow into any one junction (connection point) must also sum to zero.  If we wish to interrupt the current in a circuit, we must do so in accordance with these laws.

Although it sounds simple, interrupt the circuit, break the conduction path, or open the switch – it is not.  Forcing a conducting circuit to a steady-state condition of zero current is anything but simple.  Many times, the actual detailed physics of the process of current interruption is obscured by the seeming triviality of the switching action – such as simply flicking off a flashlight.  But consider what actually happens when a flashlight is turned off.

A steady-state direct current (DC) is flowing from the batteries to the bulb as the switch contacts begin to move.  At the last microscopic points of electrical contact, the current density becomes high enough that portions of the metallic surfaces actually melt due to resistive heating; and a liquid metal vapor plasma state continues the electrical conducting path as the contacts physically part.  As the contacts pull further apart to distances of several microns (one micron = 10^-6 meters), electrons from the contact into which the current is flowing, the cathode contact, are emitted into the intercontact space region due to thermal emission (they boil off) and field emission (they are ripped from the cathode metal by electrostatic attraction forces).

A portion of these electrons emitted from the cathode collide with air molecules within the contact gap and ionize the molecules. This frees still more electrons, which in turn ionize still more air molecules.  This self-perpetuating action is an electrical breakdown phenomenon commonly referred to as an arc.  It is the arc which enables the switch to open the circuit.  The arc forms just as the contacts part, and continues to conduct the circuit current as the contacts move further and further apart.

The voltage drop across the arc – which is proportional to the arc length and inversely proportional to the arc cross-sectional size – is in series with the voltages in the circuit loop which contains the switch.  The arc voltage grows as the arc is lengthened by the physical movement of the contacts, and the arc cross-section is diminished as the arc is cooled by contact with un-ionized air molecules.

The arc voltage in low voltage DC circuits grows at such a rate that it soon exceeds, or at least matches, the source voltage in the circuit (in a flashlight the initial arc voltage exceeds the battery voltage).  When this occurs, the circuit current is driven to zero in short order.  All circuits contain a small but finite inductance, so the current cannot be driven to zero instantaneously.  When the current does reach zero, no further arc ionization takes place, and the arc is cooled even more rapidly, since it has no energy input.  If it is cooled momentarily to such a state that it is no longer a conducting medium, then the interruption process is complete and the circuit has been opened.  It is important to remember that it is the arc that forces the current to zero.  The opening of the switch forms the arc, but it is the arc which enables the circuit to be interrupted.

A switch or circuit interruption device which is intended to open alternating current (AC) circuits has a somewhat easier chore than its DC counterpart.  In AC circuits, there is no need to force a current-zero condition.  Since the current alternates about zero already, there is a natural current-zero twice in each AC cycle.  Any arc which forms in an AC switching device does not have to be stretched and cooled to the extent that the arc voltage exceeds the magnitude of the circuit source voltage.   However, this can be done if one wishes to limit the magnitude of an overcurrent by driving it down to an unnatural current-zero.

AC currents can be interrupted at a natural current-zero, which is primarily determined by the circuit alone and practically unaffected by the presence of the interruption device.  Alternately, AC currents can be interrupted at forced current-zeros, which are imposed by the action of the interruption device.  Figure 1.3 illustrates these concepts of natural and forced current-zeros in an AC circuit.

All mechanical switches and mechanical circuit interrupting devices depend on the rapid cooling of the arc medium to open an electrical circuit.  Solid-state switches do not need an arc to break a circuit, since they supply their own conducting medium, the semiconductor material itself.  A semiconductor can conduct current only as long as mobile carriers (electrons and holes) are provided from supply or injection regions within the device.  If the injection of mobile carriers in a semiconductor switch is turned off, then the semiconductor material will revert to an insulating state and block the flow of current – that is, the semiconductor switch will turn off.

The allowable current density within a semiconductor switch is much lower than that which can safely flow in a metal contact/arc switch.  Thus, the cross-sectional size of a semiconductor switch, for equal rating devices, will always be larger than that of a mechanical switch.  Even with this disadvantage, the ease with which a semiconductor switch can be controlled, and the reliability of a device with no mechanically moving parts, portends a bright future for solid-state power switches and circuit breakers.

Overcurrents and protective devices are not new subjects.  Soon after Volta constructed his first electrochemical cell, or Faraday spun his first disk generator, someone else graciously supplied these inventors with their first short circuit loads.  Patents on mechanical circuit-breaking devices go back to the late 1800’s and the concept of a fuse goes all the way back to the first undersized wire that connected a generator to a load.

In a practical sense, we can say that no advance in electrical science can proceed without a corresponding advance in protection science.  An electric utility company would never connect a new generator, a new transformer, or a new electrical load to a circuit that cannot automatically open by means of a protective device.  Similarly, a design engineer should never design a new electronic power supply that does not automatically protect its solid-state power components in case of a shorted output.  Protection from overcurrent damage must be inherent to any new development in electrical apparatus.  Anything less leaves the apparatus or circuit susceptible to damage or total destruction within a relatively short time.  

Examples of overcurrent protection devices are many:  fuses, electromechanical circuit breakers, and solid state power switches.  They are utilized in every conceivable electrical system where there is the possibility of overcurrent damage.  As a simple example, consider the typical industrial laboratory electrical system shown in Figure 1.1.  We show a one-line diagram of the radial distribution of electrical energy, starting from the utility distribution substation, going through the industrial plant, and ending in a small laboratory personal computer.  The system is said to be radial since all branch circuits, including the utility branch circuits, radiate from central tie points.  There is only a single feed line for each circuit.  There are other network type distribution systems for utilities, where some feed lines are paralleled.  But the radial system is the most common and the simplest to protect.

Overcurrent protection is seen to be a series connection of cascading current-interrupting devices.  Starting from the load end, we have a dual-element or slow-blow fuse at the input of the power supply to the personal computer.  This fuse will open the 120 volt circuit for any large fault within the computer.  The large inrush current that occurs for a very short time when the computer is first turned on is masked by the slow element within the fuse.  Very large fault currents are detected and cleared by the fast element within the fuse.  

Protection against excess load at the plug strip, is provided by the thermal circuit breaker within the plug strip.  The thermal circuit breaker depends on differential expansion of dissimilar metals, which forces the mechanical opening of electrical contacts.  

The 120 volt single-phase branch circuit, within the laboratory which supplies the plug strip, has its own branch breaker in the laboratory’s main breaker box or panel board.  This branch breaker is a combination thermal and magnetic or thermal-mag breaker.  It has a bi-metallic element which, when heated by an overcurrent, will trip the device.  It also has a magnetic-assist winding which, by a solenoid type effect, speeds the response under heavy fault currents.

All of the branch circuits on a given phase of the laboratory’s 3-phase system join within the main breaker box and pass through the main circuit breaker of that phase, which is also a thermal magnetic unit.  This main breaker is purely for back up protection.  If, for any reason, a branch circuit breaker fails to interrupt overcurrents on that particular phase within the laboratory wiring, the main breaker will open a short time after the branch breaker should have opened.

Back-up is an important function in overload protection.  In a purely radial system, such as the laboratory system in Figure 1.1, we can easily see the cascade action in which each overcurrent protection device backs up the devices downstream from it.  If the computer power supply fuse fails to function properly, then the plug strip thermal breaker will respond, after a certain coordination delay.  If it should also fail, then the branch breaker should back them both up, again after a certain coordination delay.  This coordination delay is needed by the back-up device to give the primary protection device – the device which is electrically closest to the overload or fault – a chance to respond first.  The coordination delay is the principal means by which a back-up system is selective in its protection.

Selectivity is the property of a protection system by which only the minimum amount of system functions are disconnected in order to alleviate an overcurrent situation.  A power delivery system which is selectively protected will be far more reliable than one which is not.

For example, in the laboratory system of Figure 1.1, a short within the computer power cord should be attended to only by the thermal breaker in the plug strip.  All other loads on the branch circuit, as well as the remaining loads within the laboratory, should continue to be served.  Even if the breaker within the plug strip fails to respond to the fault within the computer power cord, and the branch breaker in the main breaker box, is forced into interruptive action, only that particular branch circuit is de-energized.  Loads on the other branch circuits within the laboratory still continue to be served.  In order for a fault within the computer power cord to cause a total blackout within the laboratory, two series-connected breakers would have to fail simultaneously – the probability of which is extremely small.

The ability of a particular overcurrent protection device to interrupt a given level of overcurrent depends on the device sensitivity.  In general, all overcurrent protection devices, no matter the type or principles of operation, respond faster when the levels of overcurrent are higher.

Coordination of overcurrent protection requires that application engineers have detailed knowledge of the total range of response for particular protection devices.  This information is contained in the “trip time vs. current curves,” commonly referred to as the trip curves.  A trip time-current curve displays the range of, and the times of response for, the currents for which the device will interrupt current flow at a given level of circuit voltage.  For example, the time current curves for the protection devices in our laboratory example are shown superimposed in Figure 1.2.

The rated current for a device is the highest steady-state current level at which the device will not trip for a given ambient temperature.  The steady-state trip current is referred to as the ultimate trip current.  The ratings for the dual-element fuse in the computer power supply, the plug strip thermal breaker, the branch circuit thermal-magnetic breaker and the main circuit thermal-magnetic breaker are 2, 15, 20, and 100 amps, respectively.  Note that, except for the fuse curve, each time-current curve is shown as a shaded area, representing the range of response for each device.  Manufacturing tolerances and material property inconsistencies are responsible for these banded sets of responses.  Trip time-current information for small fuses is usually represented in a single-value average melting time curve.

Even with a finite width to the time-current curves, we can easily see the selectivity/coordination between the different protection devices.  For any given steady-state level of overcurrent, we read up the trip time-current plot, at that level of current, to determine the order of response.

Consider the following three examples for the laboratory wiring, plug strip, and computer system.  

Example 1: Component failure within the computer power supply:  Assume that a power component within the computer power supply has failed – say two legs of the bridge power rectifier – and that the resulting fault current within the supply, limited by a surge resister, is 70 amps.

We see from the fuse trip curve that it should clear this level of current in approximately 20 milliseconds.  If the fuse fails to interrupt the current – or worse, if the fuse has been replaced with a permanent short circuit by a gambling repairperson – the thermal breaker in the plug strip should open the circuit within 0.6 to 3.5 seconds.  The branch thermal-magnetic breaker will open the entire branch circuit within 3.5 to 7.0 seconds, should the plug strip thermal breaker also fail to respond.  Note that no back-up is provided for this particular fault after the branch circuit breaker.  The main laboratory 100 amp thermal-magnetic unit would respond only if the other loads within the entire laboratory totaled greater than 30 amps at the time of the 70 amp power supply fault.

Example 2:  Plug strip overload:  Assume that the computer operator has spilled a drink, and to dry up the mess plugs two 1500 watt hair dryers into the plug strip.  The operator then flips them both on simultaneously, drawing a total plug strip load current of approximately 30 amps.

From the thermal breaker trip curve, we see that the plug strip unit should clear this overload within 5 to 30 seconds.  Note the similarity between the trip curves of the plug strip thermal unit and the branch circuit thermal-magnetic unit in the region of 100 amps and below.  This is because, for these levels of currents, the thermal portion of the detection mechanism within the thermal-magnetic branch breaker is dominant. 

Example 3:  Short circuit within the computer power cord:  Assume a frayed line cord finally shorts during some mechanical movement.  Assume also that there is enough resistance within the circuit, plug strip, and line cord system to limit the resulting fault current to 300 amps.  This level of current is 2000% (20 times) of the rated current of the plug strip thermal breaker, and is beyond the normal range of published trip time specifications for thermal breakers (100% to 1000% of rated current).  Thus the exact trip time range of the thermal unit is indeterminate.

At high levels of fault current, greater than 150 amps in this case, we can see the inherent speed advantage of magnetic detection of overcurrents.  This is evidenced by the fact that the response curve for the thermal-magnetic branch circuit breaker knees downward sharply at current levels between 150 and 200 amps.  At these and higher currents, the magnetic detection mechanism within the thermal-magnetic unit is dominant.  The response curve for the unit crosses over the plug strip thermal breaker response curve (assuming that it extends past its 1000% limit), and coordination between the two interrupters is lost.  The range of response for the thermal-magnetic breaker at 300 amps is 8 to 185 milliseconds.  Should both the plug strip breaker and the branch circuit breaker fail to operate, the main laboratory breaker should clear the fault within 11 to 40 seconds.