Small currents in HV circuitry can be measured by the usual means, resistors across which the voltage is measured or transformers. The challenge is usually in finding a way to isolate the measurement circuit so that it can "float" at the high voltage, although, with modern fiber optic techniques this is less of a problem. Often, just mounting a conventional meter in an insulating clear box is sufficient. An optical telescope can be used to read the meter from a distance.
For large, pulsed, currents however, the challenge is substantially greater. The methods essentially fall into two general categories: 1) measuring the voltage drop across a resistor; and 2) measuring the magnetic field that inevitably accompanies a current. The latter is typically done either by using a sensing coil, or more recently, by using a hall effect sensor. Another approach is to use the Faraday rotation effect to sense the field optically.
Measuring the voltage drop across a resistor (and using Ohm's Law) is often referred to as a "current shunt", deriving from the common usage where a low resistance is connected in parallel with a higher resistance meter, with the values chosen so that the current flowing through the meter is some convenient fraction (i.e. 1/100th) of the total current.
There are a number of problems with shunts in high current pulsed applications, although, the simplicity of the approach is appealing.
The resistance is typically very small (0.001 ohms would produce 1 volt/kA): you want the voltage drop to be conveniently measurable; and, you don't want to dissipate a lot of energy in the resistor (.001 ohms dissipates 1 kW at 1 kA... that Isquared term adds up..). Measuring and calibrating such tiny resistances is difficult.
The resistance varies as a function of temperature and mechanical mounting stress. Some of the temperature effects can be handled by calibration: measuring the resistance at various temperatures. However, this assumes that the entire resistor is at thermal equilibrium, which is rarely the case. As the pulse goes through the resistor, it will heat up (that kW..), and different parts will heat at different rates. Over time the heat diffuses through the structure, some radiates away, some is convectively transferred, making knowing exactly what the temperature of any part (and, hence its resistance) is very difficult. Mechanical mounting stress causes errors by changing the dimensions and contact resistances in the resistor.
Magnetic effects cause problems. One of the most troublesome is the propagation of the field into the conductor: skin effect. This can be mitigated to a certain amount by making the resistive element a loop or coaxial, and, in any event, can theoretically be calibrated out with precision measurements at low power. Another, less obvious, source of problems is deformation of the resistor by the high fields, which can stretch and bend the conductors. The series inductance of the resistor may also perturb the behavior of the system. More design information for loop shunts.
The foregoing may make it seem that anything would be better than a resistor. However, for a vast number of applications, the simplicity of resistive techniques makes them attractive. If all you need is 5% accuracy, then obsessive attention to temperature coefficients, and so forth, will be moot.
imeas.htm / 26 Sep 2000 / Back to HV Home / Back to home page / Mail to Jim