1. Introduction

In electrostatic discharge (ESD) measurements it is of interest to determine characteristics such as the current waveform and charge transferred, and peak discharge current. These are of interest in electromagnetic compatibility, and investigations into the incendivity of electrostatic discharges, and characteristic of ESD arising from metal objects, insulator surface, textiles or other materials. Until recently, electrostatic discharges have mainly been investigated by charge transfer into a reference capacitor. This method, however, does not yield information on waveforms and discharge peak current, and cannot distinguish between single and multiple discharge events. This paper demonstrates that simple probes can be fabricated using standard coaxial cable transmission lines and resistor networks, and can give acceptable performance in a wide range of high bandwidth ESD current measurements.

2. A simple transmission line ESD probe

Transmission lines are widely used conducting radio frequency signals and have wide bandwidth and low loss transmission properties. Whilst many different forms of conductors can be considered as transmission lines, a common configuration is the 50W coaxial cable. This is well known in instrumentation as a shielded cable, where the grounded outer sheath is used to reduce interference to the inner signal cable. The coaxial cable can also be used as a very simple, wide bandwidth, rugged and useful ESD probe for use in conjunction with a fast digital storage oscilloscope.

The cable is constructed such that it has a constant capacitance and inductance per unit length due to its well defined geometry and insulator dielectric properties (Jones et al 1993 pp2/23-2/26). The ratio of the voltage and current in the line is constant due to these characteristics, and is given as the characteristic impedance of the line (in ohms). In coaxial cables, 50W and 75W designs are common, with the former often used in instrumentation and rf circuits, and the latter used in television aerial leads.

Figure 1. Simple transmission line ESD measurement arrangement

 

A signal or wavefront injected into one end of the coaxial cable transmission line propagates along the cable as a voltage and current wave unless an impedance discontinuity is encountered. This could be for example a load resistor or oscilloscope input circuit. At this discontinuity point a reflection will occur, with part of the wave energy travelling back along the line, and part of the energy dissipated in the load. If the load is perfectly matched in impedance to the transmission line, then no reflection occurs and all the wave energy is transmitted into the load.

In the simple case described here, a 50 W coaxial cable is used as a 50 W impedance ESD probe. At one end of the cable the inner core is bared for a short length and may, if desired be soldered to a small metallic bead or other convenient search electrode. A correctly matched load is crucial in preventing reflections at the line termination. Many modern oscilloscopes have a 50 W input impedance option, or alternatively can be fitted with a 50 W terminator at the input socket.

The oscilloscope should have a fast single-shot digitising capability to match the expected speed of the ESD waveform. Using standard coaxial cable and a modern low cost fast oscilloscope with a sampling speed of 2 giga samples/sec, a bandwidth of 400 MHz and resolution of waveforms in the nanosecond region is feasible. Most modern oscilloscopes will allow the digitised waveform data to be downloaded onto a disk in a format suitable for import into a spreadsheet. Once in this form, the waveform can be easily plotted and calculations made on the data points as required.

A wide range of input waveform magnitudes may be investigated using this simple configuration, using the oscilloscope input attenuators to adjust the sensitivity. Care must be taken, however, not to exceed the oscilloscope input circuit ratings if larger energy discharges are investigated. Some additional attenuation and input circuit protection may be added by inserting in-line attenuators between the probe line and oscilloscope input circuit. These, however, are not normally designed to withstand high voltage or energy pulses which may result from larger discharges.

If the transmission line is correctly matched into the same impedance load, the ESD current pulse waveform is faithfully reproduced as a voltage pulse. A mismatched load may cause reflected identical or inverted waves to be seen at a fixed time interval after the initial wave, dependent on the cable length. The waveform rise time can be measured from the display in the normal way. Instantaneous or peak waveform voltage directly measured from the oscilloscope display can be converted into equivalent discharge current by a simple "ohms law" calculation using the cable characteristic impedance (in this case 50 W ) as the conversion factor.



Figure 2. Discharge from PTFE surface tribocharged using paper towel

With an oscilloscope input sensitivity of 10V/division, a current sensitivity of 200 mA per division is achieved. With an oscilloscope input sensitivity of 2mV/division, a current sensitivity of 40m A per division is possible.

If the complete discharge waveform has been digitised, then the charge transferred in the discharge may be obtained by integration of the discharge current waveform. This can be conveniently done via importing the waveform data into a spreadsheet. The total waveform charge can be obtained by summing the equivalent charge of all data points in the waveform.

Figure 2 shows a typical discharge obtained from a PTFE surface using the simple transmission line probe. The relatively slow 23ns risetime and unidirectional form of the discharge can be clearly seen, with 16nC of charge transferred in a discharge of 100ns duration and peak current of 0.13A. In contrast, a discharge from a small metal bolt is extremely fast (Figure 3). In this case 0.7nC of charge was transferred in a discharge of about 1ns duration, with a peak current of 0.82 A. The accuracy of the measurements in this case must be considered doubtful as they are limited by the 0.5 ns sampling time of the oscilloscope used.



Figure 3. Discharge from M4x15mm bolt induction charged by proximity to charged PTFE sheet

Figure 4. Current probe for monitoring capacitive discharge ESD current

3. An ESD probe using resistor elements to reduce input impedance and sensitivity

The 50 W input impedance of the simple probe may be unacceptably high for higher current discharges, and gives inadequate protection to the oscilloscope input. A network of resistors can be added to improve performance in these respects. Figure 4 shows an example we have used in monitoring discharge currents from a capacitive ESD generator circuit.

The ESD current is passed through a 0.2 W shunt resistor fabricated from five 1 W carbon resistors connected in parallel. Carbon resistors were used here as they have a wide bandwidth (Mazda 1989, p 12/10) and an exceptional capability for withstanding high peak pulse current and power dissipation. Lead lengths were kept short to reduce inductance in the ESD current path to a minimum.

The voltage developed across the shunt resistor is monitored via a divide-by-two attenuator formed by a 50 W resistor in series with the 50 W transmission line input impedance. The transmission line was terminated in a matched 50 W load at the oscilloscope input. The shunt resistor developed a voltage output of 0.2V/A, giving an overall probe sensitivity of 0.1V/A.

Figure 5 shows a discharge from a 188pF capacitor and switched into a 2.2 W resistive load using a vacuum relay. The 950nC charge is transferred with a peak current of about 200A and the waveform rings at 13 MHz. A considerable amount of VHF ringing is evident at the wavefront due to excitation of stray circuit resonances by the high initial rate of current rise.



Figure 5. ESD from 188pF capacitor pre-charged to 4kV and switched into a 2.2 W 0.7m H load

Resistor networks for use in this type of circuit must be carefully designed and constructed as stray capacitance and inductance associated with the components and circuit becomes important at higher frequencies. A typical carbon resistor may have a parallel capacitance of about 0.5 pF. The performance of a circuit can be seriously affected when the impedance of this capacitance reduce to a value approaching the resistor value. Keeping resistor values as low as possible reduces this effect. Circuits are typically housed in a small diecast box for screening against electromagnetic pickup.

4. Conclusions

The transmission line probe is very simple, rugged and easily used, and high bandwidth measurements (>500 MHz ) are easily achieved. Discharge waveform is preserved, and the wave shape, peak current, rise time, charge transferred and other characteristic measurements are easily observed. The probe sensitivity is high and measurement of small discharges are easily made. The impedance of the probe is rather high (50 W ), although this may not be a serious problem in investigations of low current discharges where the discharge impedance itself is high. Higher level discharges may exceed oscilloscope input circuit ratings and possibly cause damage to the input circuits. In a high current discharge, the probe impedance may be an important factor limiting the peak discharge current. Probes for higher level ESD currents can be fabricated using carefully designed and constructed resistor networks in combination with the transmission line. This type of circuit can be used to monitor ESD waveforms of hundreds of amps with bandwidths exceeding 100MHz.

The author gratefully acknowledges the assistance of ERA Technology Ltd and Wolfson Electrostatics Ltd in this work.