Functions / Requirements of Direct-Off-Line SMPS -- OVERLOAD PROTECTION

Home | Articles | Forum | Glossary | Books


In computer and professional-grade power supplies, it is normal practice to provide full overload protection. This includes short-circuit protection and current limits on all outputs.

The protection methods take many forms, but in all cases the prime function is to protect the power supply, irrespective of the value or duration of the overload, even for continuous short-circuit conditions.

Ideally the load will also be protected. To this end the current limit values should not exceed the specified current rating of the load by more than 20%, and the user should choose a supply rating to suit the application. This will usually ensure that the power supply, connectors, cabling, printed circuit tracks, and loads are fully protected for fault conditions.

Full protection is relatively expensive, and for small, low-power units (particularly flyback supplies) full protection is not always essential. Such units may use simple primary power limiting, and have some areas of vulnerability for unusual partial overloading conditions.


Four types of overload protection are in general use:

1. Overpower limiting

2. Output constant-current limiting

3. Fuses or trip devices

4. Output foldback (reentrant) current limiting


The first type is a power-limiting protection method, often used in flyback units or sup plies with a single output. It is primarily a power supply short-circuit protection technique.

This and the methods used in types 2 and 4 are electronic, and depend on the power supply remaining in a serviceable condition. The supply may be designed to shut down or self reset if the overload is removed.

In this type of protection, the power (usually in the primary side of the converter trans former) is constantly monitored. If this power exceeds a predetermined limit, then the power supply shuts down or goes into a power-limited mode of operation, In a multiple output unit, the power would be the sum of the individual outputs.

The power limiting action would normally take one of five forms:

A. Primary overpower limiting; B. Delayed overpower shutdown; C. Pulse-by-pulse overpower/overcurrent limiting; D. Constant-power limiting; E. Foldback (reentrant) overpower limiting


In this form of power limiting, the primary power is constantly monitored. If the load tries to exceed a defined maximum, the input power is limited to prevent any further increase.

Usually, the shape of the output current shutdown characteristic is poorly defined when primary power limiting is used on its own. However, because of its low cost, primary power limiting has become generally accepted in lower-power, low-cost units (particularly in multi-output flyback power supplies).

It should be noted that when a load fault develops in a multiple-output system, a line that has been designed to provide only a small proportion of the total power may be expected to support the full output power if it is the only line that is overloaded.

Often these simple primary power limiting systems give full protection only for short circuit conditions. An area of vulnerability can exist when partial overloads are applied, particularly when these are applied to a single output of a multiple-output system. Under these conditions, partial overloads may result in eventual failure of the power supply if they persist for long periods; hence it is better to remove this stress as soon as possible by turning the sup ply off. For this reason the delayed overpower trip technique Form B is recommended.


One of the most effective overload protection methods for low-power, low-cost supplies is the delayed overpower shutdown technique. This operates in such a way that if the load power exceeds a predetermined maximum for a duration beyond a short defined safe period, the power supply will turn off, and an input power off-on cycle will be required to reset it to normal operation.

Not only does this technique give the maximum protection to both power supply and load, but it is also the most cost-effective for small units. Although this method seems generally unpopular with most users, it should not be neglected, as it makes good sense to turn the power supply off when overloads occur. A persistent power overload usually indicates a fault within the equipment, and the shutdown method will provide full protection to both load and supply.

Unfortunately, many specifications eliminate the possibility of using a simple trip type of protection by demanding automatic recovery from an overload condition. It is possible that the user has specified automatic recovery because of previous bad experience (e.g., "lockout" or nuisance shutdowns) with reentrant or trip-type systems that did not have a sufficient current margin or a delayed shutdown. The power supply designer should question such specifications. Modern switchmode supplies are capable of delivering currents well in excess of their continuous rated value for short periods of time, and with delayed shutdown they will not "lock out" even if a shutdown system has been used.

In the delayed trip system, short transient current requirements are accommodated, and the supply will shut down only if the stress exceeds safe amplitudes for long periods.

Short-lived transient currents can be provided without jeopardizing the reliability of the power supply or having a very significant impact on the cost of the unit. It is the long-term continuous current requirements that affect cost and size. There will usually be some degradation in the performance of the unit during the high-current transient. Specified voltage tolerances and ripple values may be exceeded. Typical examples of loads subject to large but short transients would be floppy disks and solenoid drivers.


This is a particularly useful protection technique that will often be used in addition to any secondary current limit protection.

The input current in the primary switching devices is monitored on a real-time basis.

If the current exceeds a defined limit, the "on" pulse is terminated. With discontinuous flyback units, the peak primary current defines the power, and hence this type of protection becomes a true power limit for such units.

With the forward converter, the input power is a function of input current and volt age; hence this type of protection provides a primary current limit in this type of circuit.

However, this technique still provides a useful measure of power limit protection so long as the input voltage is constant.

A major advantage of the fast pulse-by-pulse current limit is that it provides protection to the primary switching devices under unusual transient stress, for example, transformer staircase-saturation effects.

Current-mode control provides this primary pulse-by-pulse current limiting as a normal function of the control technique, one of its major advantages. (See Part 3, Section 10.)


Constant input power limiting will protect the primary circuit by limiting the maximum transmitted power. However, in the case of the flyback converter, this technique does little to protect the secondary output components. For example, consider a discontinuous flyback converter for which the peak primary current has been limited, giving limited transmitted power.

When the load exceeds this limit (load resistance reducing), the output voltage begins to fall. However, since it is the input (and hence output) volt-ampere product that has been defined, as the output voltage starts to fall, the output current will increase. (On short circuit the secondary current will be large and the total power must be dissipated within the power supply.) Hence this form of power limiting is normally used to supplement some other form of limiting, such as secondary current limits.


This technique is an extension of form d in which a circuit monitors primary current and secondary voltage, and reduces the power as the output voltage falls. By this means, the output current can be reduced as the load resistance falls, preventing excessive stress on secondary components. It has the possible disadvantage of "lockout" with nonlinear loads.


Power supplies and loads can be very effectively protected by limiting the maximum current allowed to flow under fault conditions. Two types of current limiting are in common use, constant current and foldback current limiting. The first type, constant current limiting, as the name implies, limits the output current to a constant value if the load current tries to exceed a defined maximum. A typical characteristic is shown in FIG. 1.

FIG. 1 Typical V/I characteristics of a "constant-current-limited" power supply, showing linear (resistive) load lines.

From this diagram, it can be seen that as the load current increases from a low value (R1, high resistance) to its maximum normal current value (R3, median resistance), the current will increase at constant voltage along the characteristic P1- P2-P3, which are all currents and voltages within the normal working range of the supply.

When the limiting current is reached at P3, the current is not allowed to increase any further. Hence, as the load resistance continues to fall toward zero, the current remains nearly constant and the voltage must fall toward zero, characteristic P3-P4. The current-limited area is often not well specified, and the working point will be somewhere in the range P4 to P4 at a load resistance of R4.

Since the current limit is normally provided as a protection mechanism for the power supply, the characteristic in the current-limited range may not be well defined. The limit current range, P4-P4, may change by as much as 20% as the load resistance is taken toward zero (a short circuit). If a well-defined constant current range is required, a "constant cur rent power supply" should be specified.

Current limiting will normally be applied to the secondary of the power converter. In a multiple-output system, each output will have its own individual current limiting. The cur rent limits will normally be set at some independent maximum value for each output line, irrespective of the power rating of the supply. If all outputs are fully loaded simultaneously, the total loading may exceed the maximum power rating of the supply. Hence, a primary power limit will often be provided to supplement the secondary current limits. Under fault conditions, both primary and secondary components are fully protected, and the loads will all have limited currents within their design maximums at all times.

This method of current limiting undoubtedly gives the user and the supply the best protection. Not only are currents limited to values consistent with the design ratings for each line, but minimum problems occur with nonlinear or cross-connected loads. The lockout difficulties often associated with foldback limit systems are completely eliminated. Also, automatic recovery is provided when an overload is removed. Moreover, such units may be operated in parallel, the only proviso being that the current limit should be set to some value within the continuous working range. This method of protection is recommended for professional-grade supplies, although it is more expensive.


Type 3 employs mechanical or electromechanical current protection devices, and these will normally require operator intervention to be reset. In modern electronic switchmode power supplies, this type of protection is normally used only as a backup to the self-recovery electronic protection methods. Hence it is a "last ditch" protection method. It is required to operate only if the normal electronic protection fails. In some cases a combination of methods may be used.

Included in Type 3 protection methods are fuses, fusible links, fusible resistors, resistors, thermal switches, circuit breakers, PTC thermistors, and so on. These devices all have their place, and should be considered for specific applications.

Where fuses are used, it should be remembered that currents well in excess of the fuse rating can be taken through the fuse for considerable periods before fuse clearance.

Also, fuses running at or near their rated value have a limited life and should be periodically replaced. Remember also that fuses dissipate power and have considerable resistance; when used in output circuits, they will often have resistance values well above the normal output resistance of the supply.

However, fuses do have good applications. For example, when a small amount of logic current (say a few hundred milliamperes) is required from a high-current output, this may be a good application for a fuse. Clearly, it would not be sensible to design a printed circuit board and connections to withstand the high current that would flow on this low-power logic board in the event of a short circuit, and a fuse could be used in this application, providing protection without excessive voltage drop. More sophisticated protection techniques may not be justified in this situation.

Fuses or circuit breakers will also be used to back up the electronic overload protection, such as SCR "crowbar" protection in linear power supplies, in many applications. In such applications the performance of the fuse is critical, and the fuse type and rating must be carefully considered.


1. What is the normal overload protection criterion for professional-grade power supplies?

2. Give four types of overload protection in common use.

3. Give the main advantages and limitations of each of the four types of protection.

Also see: Our other Switching Power Supply Guide

Top of Page

PREV.   NEXT   Guide Index HOME