Overview:

One tube is filled with sample air and one with reference air. Sound is produced and clcked as it travels through the air in each tube. Results are expressed in terms of ratios of clock counts so that cancellation eliminate source of drift. Normalization makes all analyzers functionally identical and is performed during manufacture with air in both tubes. Electrical pulses generated as sound bursts are sent and received are used to start and stop two electronic counters driven by a single oscillator. The normalization Constant (Kn), the ratio of the accumulated counts (C1 and C2), is calculated and stored on the system disk. Here, (E) represents the systematic error function and (T) represents the effect of temperature change:

C1*E*T/C2*E*T=Kn

Operation:

After normalization, the reference tube is sealed and the sample tube connected to the sampling system. All normalized analyzers give identical results (R) when air is sampled;

C1/(C2*Kn)=R=1

The equation of state for a gas (in thermodynamic terms) relates sound velocity (V) to two physical properties; a factor (F) and molecular weight (M);

V=(F/M)^1/2

Since sound velocity and sound time-of-flight over a fixed path are inversely proportional, equations (2) and (3) are combined;

C1/(C2*Kn)=V2/V1=(F2M1/F1M2)^1/2

(F) and (M) are mole-fraction weighted averages of the component values; F_air and F_H2, and M_air and M_H2. Equation (4) is expanded by introducing and solving for mole-fraction, here hydrogen concentration [H2] in units of percent LEL (lower explosive limit). The microprocessor is programmed to evaluate equation (5) using analytical results (R) from equation (2) and the values of the physical constants for air and hydrogen;

[H2]=[M_H2+M_H2F_H2(R²)]/[M_H2(R²)(F_air-F_H2)-F_H2(M_air-M_H2)]

In one second, 50 individual results are obtained. Didital filtering allows for rapidly changing hydrogen concentration during this interval, but will not allow erroneous measurements due to a malfunction to be communicated to the system computer.