The use of rogowski coils in current measurement

Abstract

Rogowski coils are the most commonly used current sensors used in EM launcher research. They are low cost, have a wide range of sensitivity and very high bandwidth. They are also easy to install on in-situ conductors and provide isolation from the power circuit. The challenge with application is that the output signal is proportional to dI/dt, not the current. The output signal must be integrated to yield current. While this is conceptually straight forward, in practice it must be done carefully if accurate measurement of transient pulses is desired. In this paper we review the most common integration techniques used with Rogowski coils, show their strengths and weaknesses and present a technique called “hybrid integration” which we developed to overcome the weaknesses of other techniques. The simplest integration technique is a passive RC integration network, or low pass filter. While it is simple and easily implemented it suffers from droop and signal attenuation. Droop can be decreased by increasing the integration time constant but that also increases the attenuation. Passive integration is rarely an acceptable choice with the pulse durations common in EM launchers. Active integration uses an integrating operational amplifier circuit. This removes the droop/attenuation of passive integration permitting integration with no droop and high gain. The downside is that any input offset voltage present in the amplifier is integrated and produces drift in the out signal. Methods of periodically zeroing the output drift and nulling input offset have been developed with limited success. The advent of high speed digital data acquisition opened the door to direct digital integration of Rogowski coil signals. This is the most common method in use today in the EM launcher community. While the technique is easily implemented and produces qualitatively correct results, it is not without challenges. Digitizers also suffer from input offset - typically a bit or two - and this ends up being integrated into the output. EM launcher dI/dt has enormous dynamic range, very often greater than the dynamic range of the digitizer. This results in inaccurate representation of the lowest levels and/or clipping of the highest levels. Finally, as in any digital data acquisition, aliasing will occur if the signal has frequencies above the Nyquist limit of the sampling. All of these issues are present in the acquisition and processing of EM launcher Rogowski coil signals, making direct integration of the signal inaccurate and unreliable. Hybrid integration overcomes all these issues and provides an accurate and reliable measure of pulsed current. Acknowledgements: Hybrid integration was originally developed by Neil Clements at IAP Research, Inc. in the late 1980's under US Air Force sponsorship. He presented the technique at the 5 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> Symposium on Electromagnetic Launch at Destin FL in 1990. His paper was not published in the proceedings and is not widely available [1]. This paper attempts to correct that deficiency and update it to reflect the dramatic improvement in digital data acquisition since that time. Neil Clement's contribution is gratefully acknowledged.