Current measuring device

[0002] 1. Field of the Invention

[0003] This invention relates to a current measuring device for measuring electrical current waveforms using a combination of a Rogowski coil and electronic processing equipment. In particular, this invention relates to improvements in such a device whereby the high frequency bandwidth of the measurement is increased whilst still retaining the capability of measuring low frequency currents.

[0004] 2. Description of Related Art

[0005] A Rogowski coil is so named following publication in 1912 of an article titled “Die Messung der Magnitischen Spannung” (Arch Elektrotech1, pp141-150) by Rogowski W. and Steinhaus W. Its principle of operation is well known and is based on the fact that if a coil of uniformly spaced turns, wound is on a former of constant cross-sectional area, is arranged to form a closed loop, then the voltage induced in the coil at any instant is directly proportional to the rate of change of the total current passing through the loop at that instant. If means can be found of integrating with respect to time the voltage produced by the coil then the voltage obtained is proportional to the current passing through the loop. The combination of a Rogowski coil and means for integrating a voltage with respect to time thereby constitutes a current measuring system commonly referred to as a Rogowski transducer.

[0006] In practice there will be some small variation of the turns density of the coil and of the cross sectional area of the former. As a result the voltage produced by the coil will be slightly dependent on the position of the current in relation to the Rogowski coil loop. It is understood that reference in this specification to a Rogowski coil includes these practical tolerances.

[0007] The former around which the coil is wound is normally non-magnetic but it may be magnetic provided that the relative permeability of the magnetic material used is sufficiently small such that the material does not magnetically saturate when used to carry a Rogowski coil.

[0008] A Rogowski transducer has the advantages that the coil may be looped around a conductor, without the necessity of disconnecting the conductor; to provide a contactless and isolated current measurement and that large currents can be measured without magnetically saturating the transducer. A further advantage is that due to the use of non-magnetic material (which does not suffer from energy losses that increase with frequency), a Rogowski transducer potentially has a very high bandwidth, significantly in excess of 1 MHz and is therefore able to measure very rapidly changing currents.

[0009] Examples of known Rogowski transducers are described in the publications by Ray W F and Davis R M: “Wideband Rogowski current transducers: Part 1—The Rogowski Coil”, EPE Journal, Vol 3, No.1, March 1993, pp 51-59, and Ray W F: “Wideband Rogowski current transducers: Part 2 “The Integrator”, EPE Journal, Vol 3, No.2, June 1993, pp116-122.

[0010] The means for integrating the Rogowski coil voltage with respect to time may take various forms, some called “passive” means in as much as the means only utilises passive electrical components such as capacitors and resistors. Others are called “active” means in that the means also utilises active electronic components, such as semi-conducting devices and integrated circuits.

[0011] FIG. 1 shows a Rogowski current transducer that was proposed by J. A. J. Pettinga and J. Siersema in their paper “A polyphase 500 kA current measuring system with Rogowski coils”, Proc IEE, Vol 130, Pt B, No. 5, September 83, pp 360-363. This measuring system incorporates two types of passive integration, which is relevant at high frequencies, and integration using a conventional non-inverting operational amplifier, which is called “active” integration, for low frequencies.

[0012] In the circuit of FIG. 1, A represents the coil with distributed inductance L and capacitance C. The coil is connected to the remainder of the circuit by a co-axial cable terminated by a resistor R_Cof 50Ω, which is the characteristic impedance of the cable such that the terminating resistance seen by the coil is R_C. The components R₃, R₄and C₂comprise a passive integration network for which R₃>>R_Cand R₃>>R₄. A non-inverting operational amplifier circuit D acts as an integrator at low frequencies and as a unity gain amplifier at high frequencies.

[0013] FIG. 2 shows the overall frequency characteristic for the integration which falls into three bands—active integration for frequencies f in the range f₀<f<f₁, passive CR integration for f₁<f<f₂and passive L/R integration for f>f₂.

[0014] The resistors and capacitors of FIG. 1 are chosen such that for each frequency band the following behaviour occurs:

[0015] (a) The resistance R₁is relatively large and its presence is ineffective for frequencies f>f₀.

[0016] (b) For f₀<<f<<f₁the impedance of L and the admittance of C₂are negligible and the voltage V⁺ at the non-inverting input of the operational amplifier is substantially the same as the voltage E induced in the coil. For integrator gains greater than 1 the behaviour of integrator D is represented by the well known relationship: $\begin{array}{cc}{V}_{\mathrm{out}}=\frac{1}{C_1{R}_2}\int Edt& \left(1\right)\end{array}$

[0017] (c) For f₁<<f<f₂the impedance of L and C₁are negligible and the circuit D acts as a unity gain amplifier. Since the impedance of R₄is also negligible, the network R₃−C₂behaves as a passive integrator.

[0018] (d) For f>>f₂circuit D continues to act as a unity gain amplifier and the impedance of C₂is negligible compared with that of R₄. The network L-R_Cbehaves as a passive L/R integrator. This is the type of integration that has been favoured in other known Rogowski transducers.

[0019] It will be appreciated that there are significant design constraints on the relative values for the resistors and capacitors used in order to provide the required transition of behaviour from one mode of integration to the next Furthermore, in order to provide a straight-line gain-frequency relationship for the integration as shown in FIG. 2 it is important that two pairs of time constants are accurately matched, namely $\begin{array}{cc}{R}_2{C}_1=R_3{C}_2\mathrm{and}\frac{L}{R_c}=R_4{C}_2& \left(2\right)\end{array}$

[0020] The matching requirement presents difficulties both in the design and in the practical setting-up and calibration of a transducer. This is disadvantageous.

[0021] The circuit of FIG. 1 suffers from several additional disadvantages that have not until now been appreciated.

[0022] Firstly the circuit of FIG. 1 uses L/R integration. It is commonly thought that L/R integration does not result in unwanted signal oscillations. However, this is because previously published analysis of Rogowski transducers that utilise L/R integration is based on a symmetrical arrangement in which the coil loop is circular, the current to be measured lies along the axis of this circular loop and there are no other currents close to the coil. In this special case each element of the coil generates the same elemental voltage and the transit times from each element to the coil termination can be averaged to produce a smooth output voltage. The coil is therefore not susceptible to oscillations. As a result, L/R integration has been utilised in prior art transducers with satisfactory results provided care is taken to ensure a symmetrical geometry as described above. However, to utilise a symmetrical arrangement it is generally necessary to use a coil with a rigid former that cannot be opened and is therefore less convenient. Furthermore, when measuring currents in closely spaced equipment it is difficult and often impossible to arrange for the current to be central and co-axial with the coil loop. There are also often currents other than the current being measured which are close to the coil. In practice, therefore, the arrangement is not symmetrical. When this is the case, the coil is susceptible to oscillations. One solution that has been proposed is to design the coil so that its natural frequency is 4 to 5 times higher than the required bandwidth for the measured current. This solution has the disadvantage that the bandwidth of the measuring system is reduced by a factor of 4 to 5 times what could be achieved if L/R integration is not utilised.

[0023] Returning to the circuit of FIG. 1, this was designed for continuous sinusoidal currents with a bandwidth of only 100 kHz. With these specific limitations the possibility of coil oscillations (which are typically several MHz) is significantly reduced and this may enable L/R integration to be used. However for current waveforms with switching transients, such as for currents in power semi-conductors, high frequency harmonics exist which extend above 1 MHz and so cause oscillations.

[0024] In addition to L/R integration, the transducer of FIG. 1 also uses CR integration. It is known that circuits that use such integration can also suffer from the problem of unwanted oscillations One solution to this problem that has been proposed is to reduce the length of the cable connecting the coil and the CR part of the circuit.

[0025] It has only now been appreciated by the inventor of the invention described herein that high frequency oscillations in Rogowski transducers, such as those of FIG. 1, result from a mismatch between the characteristic impedance of the coil and the coil termination. This arises because the terminating resistance that is appropriate for the L/R integration of FIG. 1 is relatively small, typically 50 ohms or less, whereas the characteristic impedance of a Rogowski coil is typically 500 ohms or more. This mismatch of termination causes the output voltage to be subject to high frequency oscillations initiated by current changes in conductors outside the coil loop as well as the current being measured. Mismatch of the termination is inevitable in the circuit of FIG. 1.

[0026] A further disadvantage of the measuring system of FIG. 1 is that it does not enable both the Rogowski coil and the connecting cable to be separately terminated with the correct impedance. It will be seen in FIG. 1 that if the resistor R_Cis chosen to correctly terminate the cable then this resistance also becomes the terminating resistance for the coil. Therefore, unless the coil has the same characteristic impedance as the cable, the coil termination will be mismatched.

[0027] A yet further disadvantage of the measuring system of FIG. 1 is that the cable capacitance adds to the coil capacitance and so significantly reduces the bandwidth of the measurement system.

[0028] An object of the present invention is to provide a current transducer that overcomes the disadvantages of the prior art.

[0029] According to the present invention there is provided a current transducer comprising a Rogowski coil having an electrically conductive coil member with a coil termination, a first integrator connected across the coil termination and a second integrator connected to the output of the first integrator, the first integrator being arranged to have a substantially constant gain at relatively low frequencies and to integrate the coil output at relatively high frequencies and the second integrator being arranged to have a substantially constant gain at relatively high frequencies and to integrate the output of the first integrator at relatively low frequencies, wherein the Rogowski coil is terminated with a resistor or resistors in combination so as to provide a terminating resistance value that is approximately the same as the characteristic impedance of the coil, thereby damping the coil and reducing its susceptibility to high frequency voltage oscillations.

[0030] Since the coil terminating resistance is matched to the coil characteristic impedance, the coil inductance plays no part in the integration performed by either of the first or second integrators. This means that high frequency oscillations associated with L/R integration are avoided and the usable bandwidth of the transducer can be significantly extended. For example, the transducer in which the invention is embodied may have a bandwidth extending from 1 Hz to greater than 1 MHz.

[0031] Preferably, the first integrator is a passive integrator. Preferably, the second integrator is an active integrator.

[0032] Preferably, the passive integrator is a network of passive components. This may include the coil damping resistor.

[0033] Preferably the active integrator uses an operational amplifier in a non-inverting mode.

[0034] The frequency at which the first integrator ceases to operate as an integrator and commences to operate at a constant gain and the frequency at which the second integrator ceases to operate at a constant gain and commences to operate as an integrator may be set to be substantially the same.

[0035] The coil former may be a continuous ring made of plastic or some other suitable material and may be rigid or flexible. However, for ease of positioning the coil for use, it is preferable that the coil can be looped around a conductor without disconnecting the conductor, in which case the coil former needs to be discontinuous and of sufficiently flexible material that it can be bent into a loop around the conductor.

[0036] The passive integrator may comprise a resistor and a capacitor. An advantage of having the coil terminated with a resistance that is matched to the coil characteristic impedance in this instance is that it eliminates L/R integration and thereby avoids the difficulties of having to match CR and L/R time constants.

[0037] Preferably, the resistor and the capacitor are connected by a cable, the resistor being mounted at the coil end of the cable and the capacitor being at the second integrator end of the cable. An advantage of having the resistor at the coil end of the cable is that it avoids reducing the bandwidth of the transducer. Preferably, the resistor and the capacitor are in series.

[0038] The damping resistor may be connected at the coil end of the connecting cable such that it is in parallel with the resistor-capacitor combination of the passive integrator. Additionally, a second damping resistor may be connected between the second integrator end of the connecting cable and the capacitor of the passive integrator at that end. This second damping resistor preferably has a value substantially the same as the characteristic impedance of the connecting cable to provide a matched termination to the cable and to reduce the susceptibility of the cable to high frequency voltage oscillations.

[0039] Various devices for measuring current in which the present invention is embodied will now be described by way of example only and with reference to the following drawings, of which:

[0040] FIG. 1 shows a prior-art Rogowski current transducer;

[0041] FIG. 2 shows an overall frequency characteristic for integration associated with the FIG. 1 transducer;

[0042] FIG. 3 shows an example of the construction of a Rogowski coil;

[0043] FIG. 4 is a fragmented view of a portion of the coil of FIG. 3;

[0044] FIG. 5 is a cross section through a Rogowski coil having an electrostatic screen;

[0045] FIG. 6 is a circuit diagram of a current measuring system in which the present invention is embodied;

[0046] FIG. 7 is an example of the low pass filter shown in FIG. 6, and

[0047] FIG. 8 is a modification to the circuit of FIG. 6.

[0048] Referring to FIGS. 3 to 5, a Rogowski coil is shown having a flexible circular plastic former 11 (see FIG. 5) with a coil 10 wound in a uniform helix along its length. One end of the coil “the free end” 12 is connected to a return conductor 14 running along the centre of the former, as shown in FIGS. 5 and 6. The conductor 14 may be a continuation of the coil winding inserted through a hole running centrally along the axis of the former. The other “fixed” end 16 of the coil and the other end 18 of the said return conductor 14 are received in a socket in a ferrule 20 and are connected to a coil terminating network as will be described below.

[0049] The Rogowski coil 10 is covered with one or more layers of insulating material 22. The ferrule 20 to which the coil is attached has a second socket in which the insulated free end of the coil 12 is received to reduce the gap between the two ends of the coil. The hole is of such diameter as to provide a sufficiently close fit to hold the end 12 of the coil 10 in place. The output from the first part of the coil terminating network is connected to a coaxial cable 24, which may be doubly screened.

[0050] When the coil 10 is to be positioned around a current carrying conductor, typically, the “free end” 12 is released from the ferrule 20 socket so that the coil 10 can be looped around the conductor without having to disconnect or disturb the conductor. Once the coil 10 is looped around the conductor the free end 12 is inserted into the second ferrule 20 socket so that measurement can begin.

[0051] An electrostatic screen 26 is not generally necessary but if required it may be inserted within the layers of insulation 22 surrounding the Rogowski coil 10, provided that the screen does not constitute a closed turn around the coil 10. FIG. 5 shows the cross section of a coil 10 with a concentric screen 26 having a gap 28 along the length of the coil. This is one method of ensuring the screen 26 does not constitute a closed turn. The coil screen 26 is connected to the screen of the coaxial cable 24 or to the outer screen of a doubly screened cable.

[0052] FIG. 6 is a circuit diagram of a current measuring system that comprises four parts—a Rogowski coil 30, a coaxial connecting cable 32, a passive integrator hereinafter referred to as the terminating network of which the first part 34 is at the coil end of the coaxial cable and the second part 36 is at the other end, and a non-inverting electronic integrator 38 hereinafter referred to as the integrator. Typically, the first part 34 of the terminating network is contained within a hollow space in the end of the coil former or within the plastic ferrule 20 to which the fixed end 16 of the Rogowski coil 10 is attached. The cable 24 connects the first part 34 of the terminating network to the second part 36 of the terminating network which is situated at the input of the electronic integrator 38.

[0053] The Rogowski coil 30 is represented by its equivalent circuit comprising an induced voltage source E proportional to the rate of change of the current to be measured, an inductance L and a capacitance C. The terminal 40 of the equivalent circuit corresponds to the aforesaid fixed end 16 of the coil and the terminal 42 corresponds to the connection via the axial return conductor 14 from the free end 12 of the coil. The terminal 40 is connected to the screen of the coaxial cable and the terminal 44 is connected to the inner conductor of the coaxial cable 32. At the other end of the cable 32, the cable screen is connected to the common terminal 46 of the integrator and the cable 32 inner conductor is connected to the non-inverting input 48 of the operational amplifier. The operational amplifier is powered from positive and negative voltage sources +V_Sand −V_Swith respect to the common terminal 46 of the integrator.

[0054] The coaxial cable 32 has a distributed capacitance C_Cthat is shown as a single capacitor. A capacitor C_Amay be connected across the coil end of the cable and a capacitor C_Bin parallel with a variable capacitor C_Vacross the integrator end of the cable such that in combination these add to give a total equivalent capacitance C₀=C_A+C_C+C_B+C_V. The terminating network comprises a damping resistor R_Dconnected across the ends of the Rogowski coil 30 and, in parallel with R_D, a resistor R₀in series with the combined capacitance C₀. It is preferable that the total capacitance C₀for the terminating network other than C_Cis at the integrator end of the cable (ie C_A=0).

[0055] The Rogowski coil has a natural frequency w_B=1/{square root over ( )}(LC) radians/second which represents the upper bandwidth limit for the overall transducer. In order to avoid oscillations at or around this frequency, whilst maintaining the bandwidth, it is necessary that the combined value of the resistors R_Dand R₀in parallel is approximately equal to {square root over ( )}(L/C) ohms.

[0056] There is some flexibility of choice of the values of R_Dand R₀. It is generally preferable to make R₀substantially larger than R_Din which case R_Dis set to be approximately equal to {square root over ( )}(L/C). However in some cases, for example for transducers for measuring relatively small currents for which a relatively high sensitivity in mV/A is required, it may be desirable to minimise the value of R₀and in this case the resistor R_Dmay be omitted and the value of R₀set to be approximately equal to {square root over ( )}(L/C).

[0057] With the Rogowski coil 30 terminated with the appropriate resistance value and for frequencies up to the bandwidth limit, the output voltage E′ from the coil is substantially the same as the voltage E induced in the coil by the current to be measured.

[0058] The output voltage V₁from the terminating network is related to the coil output voltage E′ by the phasor transfer function $\begin{array}{cc}\frac{V_1\left(jw\right)}{E^{\prime }\left(jw\right)}=\frac{1}{1+j{\mathrm{wT}}_0}& \left(3\right)\end{array}$

[0059] where w is the frequency (rad/s) and T₀=R₀C₀is a time constant.

[0060] At low frequencies, for which wT₀<<1, the terminating network has unity gain. However at high frequencies for which wT₀>>1 the terminating network has a phasor gain 1/(jwT₀) which, as is well known, corresponds to integration so that at high frequencies, $\begin{array}{cc}{V}_1=\frac{1}{T_0}\int E^{\prime }·dt& \left(4\right)\end{array}$

[0061] The terminating network therefore behaves as a unity gain network at low frequencies and as an integrator at high frequencies.

[0062] The integrator circuit 38 in FIG. 6 comprises an operational amplifier, a feedback capacitor C₁connected between the output 50 and the inverting input 52, a variable resistor R₁connected between input 52 and the common terminal 46, a low pass filter network 47 in parallel with C₁and such other components (not shown) including supply decoupling capacitors and offset balancing potentiometers which are commonly used in operational amplifier circuits.

[0063] The variable resistor R₁may comprise a fixed resistor in series with a variable resistor of combined value R₁that can be varied by adjustment of the variable resistor.

[0064] To reduce the electrical noise generated by the operational amplifier at low frequencies it is desirable to utilise a low pass filter 47 in parallel with the integration capacitor an example of which is shown in FIG. 7. This filter sets the low frequency bandwidth limit w_Afor the transducer below which satisfactory integrator operation is not possible. The filter has the effect of reducing integrator gain for frequencies below w_A. Its features, together with alternative ways of achieving the same effect, arc the subject of GB 2,259,150 B and are further explained in the previously mentioned publication titled “Wideband Rogowski current transducers: Part 2 “The Integrator”” by W. F. Ray.

[0065] For some applications the measurement of currents at relatively low frequencies is not required. As a consequence the low frequency bandwidth limit w_Acan be set to a relatively higher value such that low frequency noise is not a problem. In such cases the low pass filter network can be replaced by a single resistor linking the output 50 of the operational amplifier to its inverting input 52.

[0066] Provided the frequency is greater than the low frequency bandwidth limit w_A, the phasor transfer function relating the transducer output voltage V_OUTto the input V₁to tie integrator circuit is given by $\begin{array}{cc}\frac{V_{\mathrm{OUT}}\left(jw\right)}{V_1\left(jw\right)}=\frac{1+j{\mathrm{wT}}_1}{j{\mathrm{wT}}_1}& \left(5\right)\end{array}$

[0067] where w is the frequency (rad/s) and T₁=R₁C₁is a time constant.

[0068] At high frequencies, for which wT₁>>1, the integrator has unity gain. However at low frequencies for which wT₁<<1 the integrator has a phasor gain 1/(jwT₁) which, as is well known, corresponds to integration so that at low frequencies, $\begin{array}{cc}\mathrm{Vout}=\frac{1}{T_1}\int V_1·dt& \left(6\right)\end{array}$

[0069] The integrator circuit of FIG. 6 therefore behaves as an integrator at low frequencies and as a unity gain amplifier at high frequencies, which is the converse of the behaviour of the coil terminating network.

[0070] It is now necessary to consider the relative performance at low and high frequencies. At low frequencies the integrator gain constant is 1IT₁whereas at high frequencies it is 1/T₀. If these gains are not the same, distortion of the measured current waveform will result since low frequency components will be measured at a different sensitivity compared with high frequency components.

[0071] For satisfactory measurement, typically, the time constants T₁and T₀should be equal to each other such that the low frequency and high frequency gains are the same.

[0072] This requirement is demonstrated by combining equations (3) and (5) to give $\begin{array}{cc}\frac{V_{\mathrm{OUT}}\left(jw\right)}{E^{\prime }\left(jw\right)}=\frac{1+j{\mathrm{wT}}_1}{j{\mathrm{wT}}_1\left(1+j{\mathrm{wT}}_0\right)}& \left(7\right)\end{array}$

[0073] By making T₀=T₁ $\begin{array}{cc}\frac{V_{\mathrm{OUT}}\left(jw\right)}{E^{\prime }\left(jw\right)}=\frac{1}{j{\mathrm{wT}}_1}& \left(8\right)\end{array}$

[0074] The combination of terminating network 34, 36, cable 32 and non-inverting integrator 38 will therefore provide the required integrator behaviour throughout the frequency range from the lower bandwidth limit w_Athrough to the upper bandwidth limit w_B.

[0075] The procedure to equalise the time constants T₀and T₁and to calibrate the transducer may be as follows. Using a test current of known magnitude and low frequency such as 50 Hz or 60 Hz for which the gain of the terminating network is unity, the time constant T₁of the integrator can be adjusted using the variable resistor R₁to give the desired overall gain of the transducer (for example 1 mV output per Amp of measured current). The low frequency test current is then replaced by a current with a high frequency harmonic content (for example the current in a semiconductor switch which has on and off transients with a very fast rate of change) such current waveform being separately monitored by an alternative transducer of high bandwidth such as a co-axial shunt and compared with the current as measured by the Rogowski transducer. The time constant T₀of the terminating network is then adjusted using the variable capacitor C_Vso as to make the waveform from the Rogowski transducer co-incidental and of the same shape as the waveform from the alternative transducer. When the two waveforms are the same T₀=T₁.

[0076] As an alternative to the test current with a high frequency harmonic content, a sinusoidal current at a frequency greater than 1/T₁could be used.

[0077] Various alternative arrangements to the circuit of FIG. 6 can be used to vary the time constants T₀and T₁. For example, the variable capacitor C_Vcan be omitted such that capacitance C₀has a fixed value. In order to vary the time constant T₀it is then necessary to vary the resistance R₀. This may be accomplished by inserting a variable resistor in series with R₀at the coil end 30 of the connecting cable 32. Similarly, the variable resistance R₁can be replaced by a fixed value resistor. In order to vary the time constant T₁, it is then necessary to vary the capacitance C₁. This may be accomplished by inserting a variable capacitor in parallel with C₁.

[0078] The circuit of FIG. 6 can be further adapted to include a resistor R₂in series with the capacitor C₁, so that the series combination of C₁and R₂is connected between the inverting input 52 and the output 50 of the operational amplifier. This may provide greater flexibility in the choice and setting of the time constants.

[0079] With R₂included it may be shown that $\begin{array}{cc}\frac{V_{\mathrm{OUT}}\left(jw\right)}{E^{\prime }\left(jw\right)}=\frac{1+j{\mathrm{wT}}_2}{j{\mathrm{wT}}_1\left(1+j{\mathrm{wT}}_0\right)}& \left(9\right)\end{array}$

[0080] where T₂=C₁(R₁+R₂).

[0081] By arranging for the time constants T₀and T₂to be equal in value, equation (8) applies and integrator behaviour throughout the frequency range is achieved as before.

[0082] A further modification of the circuit of FIG. 6 is shown in FIG. 8. It is standard practice with a communication cable to terminate it with its characteristic impedance by connecting a resistance of this value directly across the receiving end of the cable. This avoids the possibility of high frequency oscillations due to signal reflections from the receiving end. However, to apply this technique to the circuit of FIG. 6 would significantly attenuate the signal due to the series resistance R₀which is likely to be significantly greater than the characteristic impedance (typically 50 ohms) of the cable. In order to avoid this attenuation whilst still terminating the connecting cable 32 with an impedance which matches the characteristic impedance of the cable an additional resistor R₃is inserted between the connecting cable and the passive integration capacitors C_B, C_Vin the second part 36″ of the terminating network. At high frequencies the capacitors C_Band C_Vhave an impedance substantially lower than the resistance R₃and therefore at high frequencies R₃is effectively connected across the receiving end of the cable thereby providing the desired terminating impedance.

[0083] The value of R₃is arranged to be substantially the same as the characteristic impedance of the connecting cable, typically 50 ohms or 75 ohms. The value of R₀(see FIG. 6) is preferably arranged to be substantially greater than R₃such that, in comparison with R₀, R₃has an insignificant effect on the time constant of the passive integrator, T₀=C₀R₀, as given by equation (3). The additional presence of R₃therefore does not undermine the desired passive integration of the terminating network 34, 36 of FIG. 6.

[0084] The output of the Rogowski transducer will generally be connected to some form of monitoring or recording equipment such as an oscilloscope. The connection will generally be by coaxial cable. In order to facilitate the transmission of the high frequency components of the measurement it may be necessary to include a buffer amplifier between the output of the Rogowski transducer and such cable with appropriate termination impedances which match the impedance of the cable as is well known to communications engineers.

[0085] It will be appreciated that the current measuring systems of the invention described herein do not exhaustively cover all the possible integration networks that may be used to utilise the features of this invention and that any such alternative networks could be used.