|
How the Electrometer Triode can monitor X-rays.
An experimental portable photometer.
To achieve anything approaching accuracy in the measurement of light calls for something more elaborate than the ordinary photocell arrangement for detecting light rays. This article outlines the circuits used in colorimeters, optical density meters and similar devices, and describes the Electrometer Triode - a unique and interesting valve specially designed for the measurement of feeble currents.
The use of the photocell for quantitative measurements is not as simple as mere detection of light or darkness, where it is simply required to actuate a relay to switch in or out some external mechanism with pronounced changes, in light intensity. Simple circuits involving trigger action, which are so useful in such cases, can be modified in certain instances for the purpose of making rough measurements, but considerable elaboration is necessary if a high degree of accuracy is demanded, and in their simplest form such circuits are often unstable. There are numerous reasons for this, and the purpose of this article will be to enumerate the methods which have been successfully used in overcoming these difficulties.
Provided that sufficient light is available, the straightforward circuit, in which the photocell is connected directly in series with a battery and galvanometer, is often the best, since the relation between light and photoelectric current is linear and a measurable deflection can be obtained with a high-grade instrument for very small currents. The obvious objection is the use of an essentially delicate indicating instrument, which, of course, limits the application of the method. If a robust instrument is substituted the employment of some form of amplification becomes necessary.
Fig. 1. - A simple photocell circuit used for colour measurements in industrial processes.
The circuit of Fig. 1 has been applied to various forms of colorimeters and optical density meters, and while it is capable of no great accuracy it has been found sufficiently good to warrant its use in industry. The range over which the relation between light and current in the valve anode circuit is linear is comparatively small, though that range is adjustable by varying the components of the circuit. The photocell is the Osram type KMV6, and has an appreciable response throughout the whole of the visible spectrum, and although it has lower sensitivity to white light than the caesium type of cell it is also free from the comparatively large dark current which is frequently found in the latter type. The valve in Fig. 1 may conveniently be of the P410 or LP2 type, carefully aged and tested for low grid current and high insulation resistance. M is a Ferranti 0-1 milliammeter with 200 Ω adjustable shunt S1 and 25 Ω vernier shunt S2, capable of permitting readings to be taken over a comparatively wide range. By means of the grid bias battery G the anode current should be reduced to a low value when the photocell is dark, and the milliammeter reading then reduced to zero by adjusting the resistance R1, and the vernier resistance R2 in the circuit of the auxiliary battery B. It is convenient to test the scale relation of the meter for light incident on the cell by arranging the meter to read 100 at full illumination and then interposing screens of known absorption at two or three parts of the scale. No attempt should, of course, be made to operate the circuit from supply mains.
Avoiding Insulation Leakage
It must be remembered that the current amplification is high, and, consequently, it is obvious that, unless special precautions are taken to avoid insulation leakage, small changes in valve characteristics may take place and difficulty will be experienced in repeating observations to within 5%. Resistance coupling of several stages can be used as an alternative, but this necessitates a means of interrupting the light on the photocell at a frequency suitable for amplification in order to produce AC pulses on the input grid.
Thoriated Tungsten Filaments
Fig. 2. - An improved arrangement in which two valves are connected in a bridge circuit.
Considerable improvement on the circuit shown in Fig. 1 can be obtained by what is termed a 'null deflection method', using two valves connected in a bridge circuit as illustrated in Fig. 2. This arrangement is useful in cases where the photoelectric current is small compared with the backlash of the valve. In order to avoid extraneous disturbances the entire circuit, including batteries, should be housed in an earthed metal case, and the batteries kept permanently in circuit so that the valve may settle clown to a steady state. Valves with thoriated tungsten filaments of the type DE5 are most suitable. In using this circuit it is best to employ a plain indicator G and balance its reading to zero by adjusting the grid bias Eg. A change in this grid bias Eg corresponds to a change in photoelectric current equal to Eg/R so that if the cell has a linear relation between light and photoelectric current, the adjustable resistance, which should not be greater than 10 MΩ, may be calibrated to read directly, Elaborations of this circuit have been used in the measurement of colour temperature and light output of electric lamps, and have been standardised for precision measurements of this kind. A portable photometer of an experimental type, involving an arrangement on the lines of that shown in Fig. 2, is illustrated at the head of the article.
It may be interesting to consider more closely the reasons for lack of uniformity in valve circuits for making direct measurements. Poor insulation and grid current are the chief factors. The contributory causes of grid current are:
- Positive ions from the filament, from the anode, or from the residual gas produced by electronic bombardment. Positive ions from the filament are avoided by the use of a low-temperature filament of the oxide-coated type, and from the residual gas by a low plate voltage, i.e., less than the ionising potential of the gas. The latter condition also eliminates the possibility of positive ions from the anode.
- Electrons reaching the grid, from, the filament. This trouble can be reduced by operating the grid at a sufficiently high negative potential.
- Photoelectric emission from the grid itself. The use of a dull emitting filament and operating in complete darkness will eliminate this factor.
Poor insulation may give rise to surface leakage over the internal or external surface of the glass, and the presence of getter on the pinch of the valve will provide a further contributory cause of grid current. It is clear, therefore, that grid current may be in either direction, according to the magnitude of the factors mentioned, and the uncertainty of the value of this current is a great inconvenience in the measurement of small currents.
In order to facilitate the measurement of feeble currents, a valve of special design has been developed, and is known as the electrometer triode. Incidentally, this valve is suitable in nearly all cases as a substitute for an electrometer, i.e., for measurement of hydrogen ion concentration, X-rays, on particles, etc., and therefore offers a wide field of application in applied physics and electromedical investigations.
Fig. 3. - The electrometer triode (similar to ET1), developed by the General Electric Co., Ltd., for measuring feeble currents.
Fig. 3 shows the external appearance of the type made by the General Electric Co., Ltd., and Fig, 4 the electrode assembly, the arrangement of which is unusual.
Fig. 4.bThe electrode assembly in the electrometer triode.
The anode consists of a nickel plate mounted horizontally above the pinch, and supported by two nickel wires, one of which forms the lead-in. The filament is of the single V type and mounted horizontally above the anode. The grid is supported parallel to and above the anode by a nickel wire insulated from the pinch support wire by an elongated bead of hard glass. The valve. is fitted with a standard base, the filament and anode being brought out to the pins, to which they are connected in a normal triode, while the lead from the nickel band on the pinch which forms an internal guard ring is brought out to what is usually the grid pin. A number of turns of tinned copper wire are wrapped round the wall of the valve immediately above the base, and are fixed in position with solder. One end of this coil, which forms the external guard ring, is connected to the same pin as the internal guard ring. The grid of the triode is taken out at the opposite end of the bulb, which has an extended neck and is fitted with a screw cap.
Grid potential should be greater than -1.0 Volt, with respect to negative end of filament.
The surface of the bulb should be kept clean and dry, and the triode may with advantage be kept in an enclosure with a suitable dehydrating agent such as calcium chloride.
Fig. 5. - A series of characteristic curves obtained with the electrometer triode under varying conditions of anode voltage.
Fig. 5 shows a series of characteristic curves taken under varying conditions of anode voltage, and Fig. 6 a typical amplifier circuit. One of the most recent developments in the use of this triode consists of the combination of this valve and a photo-cell together in a third exhausted envelope, an arrangement which will probably go a long way towards overcoming the inherent difficulties of highly accurate amplifier circuits. The idea is not entirely new, since considerable work has been done on the development of photo-tubes in America, these units consisting of a photocell and thermionic electrodes in one and the same evacuated bulb. The method of combining the photocell and electrometer triode appears to have great possibilities for the future in the field of quantitative measurements.
Fig. 6. - A typical amplifier circuit.
|
