Table of contents

Spray of almost uniform drop size is formed when liquid is fed under suitable conditions on to the centre of a rotating disc and centrifuged off the edge. This method of spraying has been studied over a wide range of variables and homogeneous clouds have been produced in the drop-size range from 3 mm. to 15 μ diameter. The size of the spray drops is given approximately by the equation d=3.8 (T/Dρ)½/ω where d=drop diameter, D=disc diameter, ω=angular velocity of disc, T=surface tension of liquid, ρ=density of liquid. The spray thus formed also contains a proportion of fine satellite drops, but their smaller distance of projection from the disc enables them to be removed from the cloud when their presence is undesirable. Relatively coarse sprays are easily produced by means of a simple electric motor-driven disc. The finer spray sizes require rotor speeds up to several thousands of revolutions per second and high speed air driven "tops" have been used for this purpose. Suitable designs of apparatus are described.

The concept of dynamic impedance is applied to radiation thermocouples. An equivalent circuit is derived, in terms of which the working of a thermocouple and the factors leading to high sensitivity can be visualized. Methods are given for measuring the dynamic impedance and it is shown that these measurements lead to values for thermo-electric power, heat loss from the receiver, and time constant of the thermocouple. Expressions are given, in terms of the dynamic impedance, for the important properties of radiation thermocouples, namely the signal-to-noise sensitivity, ultimate sensitivity, noise factor and power efficiency. These properties are thus obtained in terms of parameters that can be predicted from the design, or measured in the finished instrument. The sensitivity is found without recourse to micro-measurements and therefore independent of limitations set by the thermocouple amplifier. The expression for sensitivity differs from those previously published, and the discrepancy is discussed. The way in which the ultimate sensitivity might be approached in practice is indicated. The efficiency is shown to depend on the amount of radiation falling on the receiver of the thermocouple, but to be a constant fraction of the thermodynamic limit ΔT/T. For an ideal thermocouple the fraction is one-half, for an actual thermocouple it is less by a factor which we call the relative efficiency.

Steel specimens have been examined with light and electron microscopes. The replica technique has been used for the electron microscope work, and all the specimens have been studied with both "Formvar" plastic and polystyrene-silica replicas. The micrographs are critically examined and compared.

There is an obvious gain in picture sharpness in the electron micrographs, particularly in those from silica replicas. It is found, however, that interpretation of the micrographs from plastic replicas in terms of the geometry of the specimen surfaces is more straight-forward.

Spark time lags in short gaps with tungsten electrodes were measured for different degrees of oxidation and activation, and the distributions of lags examined and compared. The results, which showed a reduction of mean lag and a narrowing of distribution following oxidation and activation, are discussed in relation to the mechanism of electron emission from oxidized cathodes for impulse and static conditions, and also for the Geiger counter.

A method is described which has been developed for the measurement of initial permeabilities of ferromagnetic wires in the microwave region and successfully applied over the range from 3 to 13 cm. wavelength to a variety of materials. The measurements are carried out by comparing the attenuations of coaxial transmission lines with the ferromagnetic specimen and with a non-ferromagnetic reference material as inner conductors. The attenuation constants are derived from observations on the input impedance of the line for different terminations as embodied in a circle diagram of impedance. The results are interpreted in the light of modern theories of magnetic dispersion and an estimate of domain sizes is made. The limitations of the theoretical treatments are indicated.

The cooling method for the determination of the specific heats of liquids is further developed. The power law of cooling, dθ/dt = kθⁿ, is shown to hold only approximately, and a new method for deducing values of dθ/dt is described. The specific heats of a number of liquid carbinols are measured over the approximate range 40-70° c.