Last issue, I discussed fire enclosures and how they prevent the spread of fire. I discussed fire containment (you need a stove-like construction), flame-retardant materials (fuel-regulated fire), smothering (oxygen-regulated fire), and automatic extinguishing (not practical). I said that the best solution to fire – any fire – is to prevent fire in the first place.
Fire results from the conjunction of the four elements of fire: heat, fuel, oxygen, and flame. If any one element is missing, there is no fire.
In the last issue, we dealt with fuel and oxygen, and we assumed we had a flame. In this issue, we’ll deal with heat sources and with the management of heat.
In the October 2016 issue, I discussed pyrolysis. You will recall that pyrolysis is the chemical decomposition of a material with increasing temperature. This is the first step in the ignition process.
Fire prevention is a simple matter of preventing pyrolysis, i.e., preventing smoke. Or, at least, preventing pyrolysis gasses from reaching ignition temperature. The old adage is: Where there is smoke, there is fire. If there is no smoke, then fire is unlikely. To prevent fire, we must prevent smoke (pyrolysis). To prevent pyrolysis, we must prevent heating the fuel material to pyrolysis temperature.
We have two choices: First, assure that the electrically-caused heating will never reach the pyrolysis temperature of the fuel material. Second, select a material having a pyrolysis temperature greater than the circuit temperature.
The second method is used in electric heaters. We know that we will have lots of heat and high temperatures. The materials used in electric heaters are, usually, metal and ceramic. Both have very high pyrolysis temperatures. The temperatures are so high that we usually say that these materials, metal and ceramic, are non-flammable materials.
The first method is to limit electrical heating such that the temperatures never attain material pyrolysis temperature. The common means for controlling abnormal electric heating is by means of fuses, circuit-breakers, or thermal cut-outs.
Actually, these are not two methods. The principle is that the electrical heating shall not cause the temperature of any fuel material to increase to the pyrolysis temperature. This is a compatibility issue of electrical heating and pyrolysis temperature. Both parameters must be considered at the same time.
Schematically, we have:
- Electrical heating.
- Thermal coupling of the heat to a material.
- Fuel material heating.
Electrical heating of a material occurs when the material is thermally coupled to the heat source. The electrical heating involves two heat parameters, temperature and energy.
The first parameter is temperature. The temperature must be greater than the sum of the fuel material ignition temperature, and the temperature drop across the thermal coupling mechanism.
The second parameter is thermal energy. The thermal energy must be sufficient to heat the fuel material to ignition temperature. If the thermal energy is too small, then the fuel material will act as a heat sink and limit the temperature rise.
I want to illustrate the two ideas of temperature and thermal energy. Consider a match. The match will readily raise the temperature of a wood shaving to ignition temperature. But, the match will not raise the temperature of a log to ignition temperature.
In both cases, the match produces the same temperature and the same thermal energy. In the case of the log, while the temperature of the match flame is greater than the log ignition temperature, the thermal energy of the match flame is insufficient to raise the temperature of the log to ignition temperature.
This illustrates that material ignition requires both temperature and thermal energy. The principle is that a small thermal energy may be able to heat a small part to ignition temperature, but not a large part. (This is a simplification, but it illustrates the principle.)
In both cases, if the match flame temperature was less than the ignition temperature of the material, then, regardless of energy, or size of the material, the material would not ignite.
We want to prevent ignition. To do so, we need only control the temperature of electrical heating. If the temperature is less than the material ignition temperature, then, regardless of energy, ignition cannot occur.
Normal operation of equipment controls electrical heating. Normal operation rarely results in excessive electrical heating and consequent fire.
Virtually all electrically-caused fires occur under circuit fault conditions which cause excessive heating. So, we are concerned with electrical heating under fault conditions. (One exception is electric heaters which, depending on proximity to flammable materials, can cause fires under “normal” conditions.)
Electrical heating is the conversion of electric energy to thermal energy. Electrical heating is expressed by:
P = I × I × R
where P is power in watts, I is current in amperes, and R is resistance in ohms. (For this discussion, we will not consider heating due to electric arcs. Rather, we will consider heating in low-voltage circuits, below 300 volts peak.) Note that there are only two parameters involved in electrical heating, current, I, and resistance, R. (Further note that electrical heating is independent of voltage.)
For electrical heating to occur, power must be dissipated in a resistance. If R is zero, then there is no electrical heating. For electrical heating, there must be a value of R greater than zero.
In evaluating a product for fire, we must look for candidate resistances in a relatively high current fault path. Often, these resistances are not discrete components, but rather are components whose resistance only comes into play under fault conditions. In other words, we need to include resistances that are negligible under normal conditions, but significant under fault conditions.
Some such resistances are wire, PCB traces, connectors, switch contacts, and wire terminations. Each of these is assumed to be zero during normal operation. However, under fault conditions, with maximum current, cross-sectional area of a wire or PCB trace may be too small for the fault current and thereby overheat. Connectors and switches have contact resistance which can also dissipate power under overcurrent conditions.
Furthermore, candidate resistances must also be robust. That is, they must be capable of dissipating high power for enough time to raise the temperature of the fuel material to ignition temperature.
The other factor in power dissipation is fault current. What is the value fault current that can be expected? If the current is limited by a fuse, then the maximum continuous fault current is 110% of the fuse rating. If the current is not limited by a fuse, then the maximum fault current is determined by connecting a variable load to the circuit and adjusting for maximum current.
Maximum fault current is not the maximum rated output of a power supply. Most power supplies will output much more than rated current into fault. And, they will do so for an extended period of time. So, you must always measure the fault current with a variable load. Then, using that current, you can evaluate the various candidate power-dissipating resistances to determine if the temperature exceeds material ignition temperatures.
One caveat: Sometimes maximum fault current will cause an immediate failure of a candidate power-dissipating resistance. If this occurs, then you need to re-test at a lower value of current to evaluate the long-term dissipation. This is because not all faults are short-circuit or maximum current. A fault current can be any value exceeding rated current up to maximum fault current.
If overheating occurs, then fire prevention is a simple matter of taking steps to reduce the resistance of the power-dissipating device. Or, fire prevention is a simple matter of installing a fuse to limit the current to a value which will not cause overheating.
Fire prevention in electronic products is not a simple matter of following the construction requirements of standards. Fire prevention requires an understanding of fault currents and their paths, and power-dissipating resistances.
Fire can be prevented.
This discussion just gives a broad overview of general principles of electronic product fire causation and prevention. There is a lot more that can be said and can be researched. Even then, protection against ignition for any particular product will require more engineering and testing than currently required by our various safety standards. The extra effort will pay off in fewer fires than we now incur.
Some of the material presented here is the result of a collaboration of Dave Adams, Ray Corson, Kevin Cyrus, Richard Pescatore, and Brady Turner, all of Hewlett-Packard, and Bob Davidson and Don Mader of Underwriters Laboratories.
Richard Nute is a product safety consultant engaged in safety design, safety manufacturing, safety certification, safety standards, and forensic investigations.