If there is one problem in electronics that really gets people worried, it is fires. Fires in electronic systems naturally cause apprehension. Will it happen again? What might the consequences be in terms of personal safety and company reputation?

Where to start
The main aspects that concern customers are:

	Finding the prime cause of the fire
	Determining whether it has been caused by wear-out, misuse, or poor design
	Estimating how high a temperature was reached at various points
	Examining whether the fires are self sustaining
	Measuring whether the resulting fumes are toxic or if there are any other hazardous effects
	Assessing how likely there is to be a repetition

The first action is often to find out how it happened or whose fault it was. It might be thought that if a piece of board has a 1-3 cm diameter hole burnt in it, that it is impossible to find the

cause of the fire. Figure 1 shows an example of this type of fault. The component causing the fire was in the middle of this hole and has completely disappeared but it was still possible to tell how it happened. The techniques are based on measurements on the properties of the remnants and on our knowledge of electronic components, assembly methods and wear-out mechanisms. The causes are often a combination of factors, where each on their own is not serious.
		Figure 1 A small fire left this hole in a printed circuit board

Determining the cause
The peak temperatures reached and the thermal history of some types of plastic parts can be determined in the laboratory. For this investigation it is an advantage to have an undamaged sample for comparison.

The addition of flame retardants has drastically reduced the spread of fires but cannot be relied on to prevent fires. These materials operate by various mechanisms such as limiting the accessibility of the air required to sustain a fire. Whichever type of fire retardant is used it cannot totally prevent fire if the energy density or the temperature is high enough to totally carbonise the plastics. These then become electrically conductive and leakage currents through these extra paths can allow the fire to continue.

Figure 2. A polypropylene capacitor after a small fire		There are some situations where flame retardants cannot be used. These include the windings for plastic film capacitors, which use high purity material without any additions. The result of a breakdown and small fire is shown in Figure 2. The potential hazards are most serious when the equipment is close to the general public. Fortunately there are sensitive techniques available to help with the situation. The fumes given off in the fire can be analysed and compared to standard lists of the toxicity
of various substances in terms of the likely exposure time. These tests demonstrate a perhaps surprising result; intense fires can be advantageous since many of the complex and dangerous substances given off by a fire are decomposed by the high temperatures developed.
It might be thought that the prime sources of heat for causing fires are those components that dissipate the most heat, such as power resistors or transistors, but in our experience these are not often the cause of the problems. Arcing is one of the more frequent causes. Solder joints can fail by several means and it is found that fires often start here. A typical early phase is shown in Figure 3, where a crack is forming around a pin mounted through a hole in the board.		Figure 3. This crack at a solder joint is typical of many that cause small fires
Eventually arcing can cause small amounts of damage as in Figure 4, where part of a connector has been burnt away. A cross section of the neighbouring pin is shown in Figure 5, where there are severe cracks through the solder joint. This damage shows an early phase of the fault that caused the fire.
Figure 4. The remnants of a connector after a small fire		Figure 5. A cross section of the pin next to that in Figure 4, showing cracks in the solder at the plated through hole
Wet electrolytic capacitors are subject to various damage mechanisms. The electrolyte can leak out and cause conductive paths across the PCB. In other cases there is gas evolution, where its rapid expansion can causes cases to split. Another factor to bear in mind is that the electrical breakdown of a capacitor can dissipate significant energy in a short time due to the energy stored on the electrodes.

An unusual form of damage is shown in Figure 6. There were several pairs of blisters at plated through holes. By cross sectioning it was found that a layer of the printed circuit board (PCB) between these points had become charred. Careful polishing and analysis of the surfaces confirmed that the fault was due to “conductive anodic filaments”, where the matrix of the PCB separates from the glass fibres, allowing moisture and ionic contaminants to form a leakage current. The current increased sufficiently to carbonise the plastic, further increase the electrical conductivity and cause small fires.

Figure 6. Two small damaged areas that were linked by conductive anodic filaments through the board