The sensing element of a CFCG indicator is usually a catalytic metal filament heated by an electric current. In some instruments the filament is replaced by a ceramic pellet with a catalyst (a pellister) but the mode of action is the same. When a mixture of hydrocarbon gas with air is drawn over the filament the gas oxidises on the hot filament and makes it hotter. This increases its resistance and the change of resistance provides a measure of the concentration of hydrocarbon gas in the mixture.
A simplified diagram of the electrical circuit of a CFCG indicator is shown in Figure 18-1. It is a Wheatstone bridge with the sensor filament forming one arm of the bridge.
The indicator is made ready for use by balancing the bridge with the filament at the correct operating temperature in contact with fresh air, so that the meter reading is zero. The increased resistance of the sensor filament brought about by combustion of the sample mixture throws the bridge out of balance and causes the meter to deflect by an amount proportional to the combustible gas concentration. The deflection is shown on a scale calibrated to read 0%-100% LFL. Some instruments incorporate additional circuitry to give a second, expanded, range 0%- 10% LFL. To maintain consistent readings the voltage across the bridge must be kept constant and a control is provided for this purpose.
Another arm of the bridge consists of a second filament (compensator filament) identical with the sensor filament and the two are mounted close to each other in the instrument. The second filament, however, remains permanently in contact with pure air and the arrangement provides
reading. The resistances in the other two arms of the bridge are made from an alloy, the electrical resistance of which is practically independent of temperature.
In taking a measurement the manufacturer’s detailed instructions should be followed. After the instrument has been initially set at zero with fresh air in contact with the sensor filament, a sample is drawn into the meter by means of a rubber aspirator bulb or a pump. The reading is taken when the pointer has ceased to rise on the scale.
The out-of-balance voltage across the meter is proportional to hydrocarbon concentration up to 2-3 times the LFL although a reading cannot go beyond 100% LFL. If the concentration is more than about twice the LFL there is insufficient oxygen in the mixture to burn the hydrocarbon gas completely. The response of the instrument to such a concentration is that the needle initially deflects to the maximum on the scale and then falls back to a reading near zero. Continuous observation of the needle is necessary to avoid overlooking this kind of response. Prolonged operation with such a gas mixture causes the deposition of carbonaceous matter on the sensor filament and may alter the response of the instrument. In such cases the response of the instrument should be checked.
For the same reason the instrument does not give a reliable reading with atmospheres deficient in oxygen, such as those in inerted tanks. The meter must not be used for measuring hydrocarbon concentrations in inerted tanks.
Non-hydrocarbon gases, such as hydrogen sulphide or carbon monoxide, or gases from lead compounds, which may be present in a tank atmosphere can affect the meter response, but only if they are present in very high concentrations.
The instrument is normally fitted with a filter to remove solid particles and liquid. It will not indicate the presence of combustible mists (such as lubricating oils) or dusts.
Figure 18-1 Simplified Circuit Diagram of a Catalytic Filament Combustible Gas
18.2.2 Instrument Check Procedures
The instrument is set up to read correctly in the factory using a hydrocarbon gas/air mixture, the composition of which should be indicated on the label fixed to the instrument.
The response should be checked at the beginning of every day during which it is intended to use the instrument. Such a check should also be made after replacing a filament. Test kits for use in the field are available for this purpose providing a mixture of hydrocarbon gas in air (such as 50% LFL butane in air). At intervals, the instrument should be checked more thoroughly in a laboratory equipped with suitable gas blending facilities.
During operation it is important to check the instrument and sample lines occasionally for leakage, since the ingress of air would dilute the sample, giving false readings.
Leak testing may be achieved by pinching the sample line and squeezing the aspirator bulb; the bulb should not expand as long as the sampling line is pinched.
18.2.3 Precision of Measurement
The response of the instrument depends upon the composition of the hydrocarbon gas and in practice this composition is not known. The calibration of the instrument is such that the response is usually on the safe side for the gases encountered in tanker operations.
Factors that can affect the measurements are large changes in ambient temperature and excessive pressure of the tank atmosphere being tested, leading to high flow rates which in turn affect the filament temperature. To avoid the effect of gas flow rate, it is recommended practice to take a reading when there is no flow, i.e. between two squeezes of the rubber aspirator bulb. The use of dilution tubes which enable catalytic filament indicators to measure concentrations in over rich hydrocarbon gas/air mixtures is not recommended.
18.2.4 Operational Features
Only instruments fitted with flashback arresters in the inlet and outlet of the detector filament chamber should be used. The arresters are essential to prevent the possibility of flame propagation from the combustible chamber; a check should therefore always be made that they are fitted properly in their place.
Some authorities require, as a condition of their approval, that PVC covers be fitted around meters with aluminium cases to avoid the risk of incendive sparking if the case impacts on rusty steel.
When hydrocarbons are being measured no filters should be used other than the cotton filter inserted in the gas inlet of the detector to remove solid particles or liquid from the gas sample, although a water absorbent material or water trap may be necessary in the sampling line if the gas is very wet (see Section 18.10).