Basic Fuse Construction and Operation

(The fuse element is made of zinc, copper, silver, aluminum, or alloys to provide stable and

predictable characteristics)

The typical fuse consists of an element which is surrounded by a filler and enclosed by the fuse body.
The element is welded or soldered to the fuse contacts (blades or ferrules).

Figure 1. A typical fuse

The element is a calibrated conductor. Its configuration, its mass, and the materials employed are
selected to achieve the desired electrical and thermal characteristics. The element provides the current
path through the fuse. It generates heat at a rate that is dependent upon its resistance and the load
current.
The heat generated by the element is absorbed by the filler and passed through the fuse body to the
surrounding air. A filler such as quartz sand provides effective heat transfer and allows for the small
element cross-section typical in modern fuses. The effective heat transfer allows the fuse to carry
harmless overloads. The small element cross section melts quickly under short circuit conditions. The
filler also aids fuse performance by absorbing arc energy when the fuse clears an overload or short
circuit.
When a sustained overload occurs, the element will generate heat at a faster rate than the heat can be
passed to the filler. If the overload persists, the element will reach its melting point and open. Increasing

the applied current will heat the element faster and cause the fuse to open sooner. Thus fuses have an
inverse time current characteristic, i.e. the greater the overcurrent the less time required for the fuse to
open the circuit.
This characteristic is desirable because it parallels the characteristics of conductors, motors, transformers
and other electrical apparatus. These components can carry low level overloads for relatively long times
without damage. However, under high current conditions damage can occur quickly. Because of its
inverse time current characteristic, a properly applied fuse can provide effective protection over a broad
current range, from low level overloads to high level short circuits.

Semiconductor Fuses - General

Figure 1. Possible arrangements for semiconductor fuses in a converter bridge (a1 and b1) and for an

AC controller (a2 und b2); a) Arm fuse or b) Phase fuse

Diodes and thyristors do not limit short-circuit current. For this reason, external measures have to be
taken to ensure short-circuit protection. The silicon chip of a thyristor or rectifier diode has a very low
thermal capacity, which is why strong, fast rising overcurrents under short circuit conditions can destroy
this part within just a few milliseconds. For this reason, normal low-voltage fuses, including quick-blow
fuses, are not suitable for protecting rectifier diodes and thyristors from being destroyed by short circuits.
Instead, purpose-developed fuses which are available as fast-blow, super-fast blow and ultra-fast blow
fuses or simple semiconductor fuses have to be used.
The most common causes of short-circuits in converters are:
Short circuit in the load or the connecting lines between the converter unit and load
Short-circuit in a rectifier diode or thyristor due to loss of blocking ability (sudden failure)
Loss of stability in inverters or converters due to trigger error
If a semiconductor fuse is connected in series to each rectifier diode and each thyristor, where the diodes
and thyristors are used in the main legs of the converter circuit, the diodes and thyristors can be
protected in all of the above short-circuit cases. At the same time, all other components in the main
circuit are also protected, since chokes, transformers, resistors, etc. are far less sensitive to overcurrents
than the semiconductor components. In bridge circuits, two power semiconductors each have a common
AC terminal. For this reason they can be protected by a common fuse in the AC supply line (Figure 1b).

This is known as a phase or AC side fuse. This has the advantage that fewer fuses are needed and that
the switching voltage that occurs when the fuse element melts does not put any load onto the
semiconductor components.
For high current loads on thyristors or rectifier diodes, as may be the case in systems with forced cooling,
and high operating voltage at the same time, it may well be difficult to find a fuse with sufficient rated
current (effective current in the phase 2 times higher than in each of the two arms) and with an
operating it, on the other hand, that is lower than that of each of the two semiconductor components.
In this case, each bridge arm must be allocated its own fuse (Figure 1a). In AC controllers, a common
fuse normally suffices (Figure 1b2).
If rectifier diodes or thyristors are connected in parallel for high power densities, each semiconductor
component is equipped with a fuse. The advantage of this is that if one component or fuse suddenly fails,
the device will remain fully functional with less power, accordingly. The internal resistances of the fuses
will also compensate somewhat for the differences in the characteristics of the semiconductor
components, meaning the current distribution will be more homogenous. For short-circuit protection
using blow-out fuses, short circuits are assumed to seldom occur. There are, however, applications where
load short-circuit happens more frequently. In this case, a DC side high-speed circuit breaker, a power
circuit breaker, or another device to block the driver pulses of the thyristor (driver signal blocking) is
needed on the load side. This has to be used in addition to the fuses to cater for the other possibilities of
short circuiting.

Basic Considerations for Semiconductor Protection

with Fuses

Solid state devices have progressed through several generations of sophistication since their introduction
in the 1940s. Fuse designs have changed to match solid state protection demands. The protection task
seems simple - choose a fuse of correct voltage and ampere rating which will protect a solid state device
(diode, silicon-controlled rectifier, triac, etc.) through a wide range of overcurrents, yet carry normal
rated loads without deterioration through a long life.

Considerations when using fuses for protection

Solid state power devices operate at high current densities. Cooling is a prime consideration. The fuse
should be cooled with the solid state device. Cycling conditions must be considered. The ability of solid
state devices to switch high currents at high speed subjects fuses to thermal and mechanical stresses.
Proper fuse selection is mandatory for long-term reliability.
Solid state devices have relatively short thermal time constants. An overcurrent which may not harm an
electromechanical device can cause catastrophic failure of a solid state device. Many solid state devices
have an overcurrent withstand rating which is termed I 2t for fusing. These values are found in most
power semiconductor application handbooks. Fuses intended for solid state device protection are rated in
terms of total clearing I2t. Fuses and devices are matched so that the total clearing I2t of the fuse is less
than the withstand I2t for the device.

Figure 1. Example of It curve for fuses

The published fuse total clearing I2t values are derived from short-circuit test oscillograms of the fuse
under controlled conditions. The end application can vary significantly from the tested conditions. The
specifier must take these differences into account since they will affect fuse clearing I2t.

Cooling Methods for Power Semiconductor Devices

Cooling methods can be classified according to the mechanism or medium used to transfer the heat
during the cooling process. A commonly used method of cooling power semiconductors is air cooling,
which includes natural air cooling and forced air cooling. Another type of cooling is liquid cooling. Liquid
cooling is often accomplished by use of water or a water/glycol mixture to perform thermosyphon cooling
or forced cooling. Other agents can also be used for liquid cooling such as oil and several other inert
fluids. Cooling can also be achieved by taking advantage of the heat transfer that occurs when materials
experience phase transitions.

Air Cooling
Air is not an outstanding thermal conductor (air has a thermal conductivity of 0.026 W/mK). There are
however advantages in using air as a cooling agent which include its universal availability, its ability to
insulate, as well as its non-corrosive nature.

Natural Air Cooling

Figure 1. Natural air cooling

It is common knowledge that air rises as it is heated due to its resulting decrease in density (convection).
The air flow resulting from this convection process is referred to as laminar flow. This process provides a
natural means of removing heat generated by power electonics components.
Advantages of applying natural air cooling include low to no maintenance requirements, no resulting wear
and tear, and no noise emission during application. The most prevalent disadvantage of using natural air
cooling is that it results in very low levels of cooling.

Forced Air Cooling

Air blowers/fans are used in forced air cooling in order to increase the air velocity. This increased velocity
aims to produce turbulent air flow rather than a laminar flow, effectively increasing heat disipation to the
surrounding atmosphere.
The advantage of using forced air is that it has a far better cooling effect than natural air cooling.
Disadvantages include the incredibly high amount of noise produced during application as well as
resulting wear and tear.
Axial fanning and radial fanning are two types of forced cooling named and classified according to how
the fan is mounted in relation to the device to be cooled.

Figure 2. Forced cooling using an axial fan

Figure 3. Forced air cooling using a radial fan

In the examples above, the insulated structure of the modules makes it possible to connect all the
semiconductors to the same heat sink. This results in a clear structure with simple bus-bars.

Liquid Cooling
A more effective method of cooling than air cooling is liquid cooling, which normally involves water or a
water/glycol mixture as the heat transfering medium.

Turbulent Flow to Improve Thermal Efficiency

One way of increasing the thermal efficiency of a water cooling system is is by placing coils inside the
cooling channel to induce a turbulant flow of the cooling liquid.

Figure 4. Use of coils in liquid cooling system

The turbulent flow created by the coils can increase the thermal efficiency of the cooling system by 15 20%.
Turbulent flow can also be created by the presence of microchannels in the cooling channel.

Figure 5. Use of microchannels in liquid cooling system

A disadvantage of this method is the high risk of channel blockage by particles in the cooling liquid.

Thermosyphon Cooling
In thermosyphon cooling, the transfer of heat is accomplished by the natural convection of water due to
gravity. This results from the fact that heated water is less dense than the cooler water and therefore
rises to the top of the cooling system causing natural circulation of the cooling liquid.

Figure 6. Thermosyphon cooling system

Advantages of using thermosyphon cooling is that a minimum amount of maintenance is required, no
wear and tear is caused, and the cooling process produces no noise. The main disadvantages of this
method is that the system must always be positioned in a vertical direction and requires a large amount
of space.

Phase Transition Cooling

The enthalpy of vaporization is the amount of heat that must be absorbed by a given quantity of liquid in
order to transition to the gas state. The opposite of this is referred to as enthalpy of condensation. The
same amount of heat is used up or dissipated in each process respectively.
The cooling fluid evaporates at the location of the heat source. The vapor carries the heat to a condensor
(which acts as a heat exchanger), where the fluid is then condensed back to its liquid form. The enthalpy
of vaporisation of cooling liquids is high (>2000 kJ/kg).
Examples of application of this method of cooling include pool boiling, heat pipes, spray cooling, jet
impingement cooling and vibration induced droplet atomisation (VIDA).

Pool Boiling
In pool boiling, the cooling medium evaporates at the heat source, gas bubbles rise and condense on the
cooler upper surface.

Figure 7. Pool Boiling

At high heat stream density, a layer of vapor might build up at the heat source. This reduces or prevents
contact between the heat source and the cooling fluid which inevitably leads to a great reduction in
cooling. This is referred to as the Leidenfrost effect.

Heat pipes
A very useful method of transfering heat away from semiconductor devices is the use of heat pipes. Heat
pipes also rely on natural forces to transfer heat. Heat pipes are made of hermetically sealed copper filled
up with a small amount of fluid under low pressure. The inner part of the heat pipe is lined with a
capillary-structured wick.

Figure 8. Liquid cooling with heat pipes

The cooling liquid is evaporated by the heat source at one end of the heat pipe. The vapor is transferred
to the opposite end by convection where cooling fins are located. The vapor cools and condenses into
liquid form and is carried back to the heat source through the capillary wick structure along the perimeter
of the heat pipe.
To form the capillary structure of the heat pipe, a porous material is applied on the inner wall of the pipe.
This can be done using either metal foams (such as steel, aluminium, copper or nickel) or using carbon
fibres. Methods of creating the capillary structure include:
Using sintered powder (sintered wick) - offers the greatest cooling effect

Increasing the surface of the inner wall (grooved tube) - very weak capillary action
Screen mesh - most often used .
Different heat transfer media can be used within the heat pipes. The choice of medium depends on the
application, in particular the required temperature range. Water, for instance, cannot be used below 0C.
Acetone or alcohol are commonly used.
Advantages of using heat pipes include:
Extremely high heat transfer ability (100 to 1000 times higher than copper at small temperature
gradients
No parts need be moved mechanically, hence requires no maintenance
Heat pipes offer enough flexibility to be produced in all forms and sizes

Spray Cooling
Spray cooling and jet impingement cooling can be carried out either from the bottom or top or from both
sides of the heat source. The cooling fluid vaporises upon coming into contact with the semiconductor
chips and condenses when it reaches the cooler areas. Fluids used for spray cooling are mostly inert
fluids such as fluorinert or other fluorinated hydrocarbons which cover a wide boiling point range.
Conductive water can only be used when spraying on the bottom side of the module.

Figure 9. Spray cooling system

A great advantage of spray cooling is the fact that, upon spraying, the cooling fluid vaporizes directly on
the chip. This leads to faster heat transfer. A big disadvantage, on the other hand, is the less than
optimal thermal characteristics of the inert fluids. As an example, table one below shows a comparison of
the thermal characteristics of fluorinert and water.

Comparison
Latent heat of vaporisation
Heat capacity

Fluorinert (FC - 72)

88 kJ/kg
1.05 kJ/kg.K

Table 1. Comparison of Thermal Characteristics of Fluorinert and Water

Water
2250 kJ/kg
4.18 kJ/kg.K

Another disadvantage of spray cooling is high technical complexity. An large amount of equipment
including pumps, spray pipes, condensors, storing tanks, microfilters, and a closed pressure system are
required in order to carry out spray cooling. The process also requires very high pressures ranging from 3
bar to 15 bar. Wire bonds on the chip are also always in the way of the sprayed fluid, hindering optimum
cooling of the chip. Also, due to their small diameter, there is always a chance that the spray pipes might
become clogged. Cavitation of the pressure pump is also quite likely to occur.

Vibration Induced Droplet Atomization (VIDA)

Figure 10. Vibration Induced Droplet Atomisation (VIDA)

In vibration induced droplet atomization, the vibration of a Piezo membrane atomizes the cooling fluid.
This atomized fluid then vaporizes at the hot areas under the chip. The vapor transfers the heat to the
cooler walls of the cell where the fluid condenses and flows back to the Piezo membrane. A great
advantage of VIDA is the high degree of cooling it provides and its very simple structure. On the other
hand, VIDA is highly dependent on positioning.

Thermal-Acoustic Cooling
Thermal-Acoustic cooling is based on the principle of a Stirling machine that uses a loud speaker as a
drive. The working medium is usually an environmental friendly rare gas such as Helium or Argon. This
heat pump is powered by very high acoustic pressure caused by resonant vibrations of the gas in the
inner and outer board. Two cooling circuits (hot and cold) for heat transfer are required to complete the
system. The advantage of using thermal-acoustic cooling is that there are no moving parts. Thermalacoustic cooling however has lower than desirable effectiveness.

Cooling using the Peltier Effect

In this process, current flows through the connecting point between two different kinds of conductors
(metal or semiconductors such as Bismuth Telluride Bi 2Te3 ). This has either a heating or cooling effect
depending on current flow direction (reversed Seebeck effect). Heat transfer occurs via electrons in nsemiconductors and via holes in p-semiconductors.

No noise is produced while cooling using the Peltier effect and the equipment used in this method
requires no maintenance. Cooling using the Peltier effect, however, has very low cooling effect.
In conclusion, adequate cooling plays an important role in the reliability of power semiconductors and
their lifetime. Choice of material as well as the structure of semiconductors determines the level of
thermal resistance of any given semiconductor device. A wide range of processes with very different
levels of effectiveness can be used for heat transfer in semiconductors.