Back to Westbay Tools

Click on the following bookmarks for further details of a Transient Tool for emc analysis:

Transient voltage clamping
Transient voltage reduction by series inductor
Charged capacitor discharge
Suppression filter step response
Analysis of hybrid transient protector
Analysis of the thermal noise in a resistor

Transient voltage clamping

The  voltage waveform utilised by this tool has a linearly rising edge followed by an exponential decay, and a defined source resistance.

    The circuit arrangement allows the above source voltage to be applied to a resistive load through a series resistance.. The load can be shunted with a voltage clamp device, which is open circuit until its threshold voltage is reached. At threshold, it maintains a constant voltage across itself while the source voltage is capable of maintaining the threshold. The clamp also has two series elements, an inductor and resistor. These can be used to make realistic estimates of the effect of lead inductance and pcb track impedances.
    The data entry form is shown below:

voltage Clamp form.gif (10165 bytes)

The circuit arrangement is shown below:   

Pulsecct.gif (3098 bytes)

    The analysis provides total energy, peak power, peak current and peak voltage in the load and series resistor. The clamp analysis provides total energy, peak power and peak current. The load voltage can be plotted, and an example plot is shown below.

Clamp_g.gif (6967 bytes)

    The load voltage when a 1500V peak voltage with a 500ohm load is clamped by a 36V clamping device. A poor ground with 50mohm resistance raises the peak voltage to 50V. The rise time of the pulse is 1us, and the decay time 5us.

Transient voltage reduction by series inductor

The waveform shown on the data entry form below covers many requirements, consisting of a linearly rising edge, followed by an exponential decay.

Ind_form.gif (9907 bytes)

    The waveform is defined by the tool in terms of the risetime to 50% of the peak value, and the time from the pulse onset until decay to 50% of the peak. During analysis, the time constant of the decay is displayed for information. The circuit arrangement used by the inductor transient attenuator is shown below:

Pulseind.gif (2318 bytes)

    A spot analysis is produced of peak current, voltage and power in the series resistor and load resistor, and also of total energy. Peak voltage and current at the inductor are also presented. A plot can be produced, which graphs the load voltage and inductor voltage against time.The performance of an inductive transient attenuator varies for a given inductance according to the pulse shape.

Ind_graph.gif (7287 bytes)

    For instance, an analysis of a 500V peak transient, of 8us rise and 20us decay times, shows a peak voltage of 48V at the load. Increasing the decay time to 200us, increases the peak load voltage to 86V, and also increases the total energy delivered to the load from 21.5mJ to 141mJ. The plot above shows the 200us decay time pulse, with the blue curve showing the inductor voltage, and the black curve the load voltage.

Charged capacitor discharge

The charged capacitor discharge tool uses the circuit arrangement shown on the form below:

Capdischarge_form.gif (7712 bytes)

    The capacitor C1 holds the initial charge, and the resistor Rs can represent both source resistance, and any additional series resistance the user may wish to add. Rl represents the load resistance, and C2 the charge balancing capacitor.

    In the example figures shown on the form above, an ESD-like test has been simulated. A 1500V charge on a 150pF capacitor has been discharged through 150 ohms into a 50 ohm load. Without the charge balancing capacitor C2, the peak voltage at the load would have been:

        VPeak = (1500 * 50) / (50+150) = 375 volts.
    Use of the 1nF capacitance has reduced the peak to 109 volts, and a larger capacitance would reduce it further. The example plot below shows the same arrangement as above, but with C2 increased to 10nF.

Capdischarge-graph.gif (7048 bytes)

    The peak voltage has been further reduced to less than 20 volts, albeit at the expense of increased capacitance on the line.

Suppression filter step response

    EMI Suppression Filters consist of shunt capacitances and/or series inductances. Their general response to the application of a step voltage is to produce a delayed rising edge at a resistive load.

    This effect can be of particular importance when the filtered line is carrying a fast digital signal. If the signal risetime at the load exceeds the period of the signal, the peak signal voltage is not reached, and the signal is effectively attenuated.

    A second effect occurs when a threshold voltage is important; the threshold is delayed by the filter, and component tolerances in the filter may cause different signal lines to produce staggered thresholds.

    The tool allows six different filter circuits to be selected, which can be shown on-screen via the Diagram button. The circuits are Capacitor, Inductor, L-C Filter, C-L Filter, Pi Filter and T Filter. Source and Load Resistance can be specified, and Inductors can also have a series resistance. The Filter Step form is shown below:

Stepfilter-form.gif (5455 bytes)

Application of the step then shows the load response; expand the timescale until the rising edge can be seen. Comparison of the edge with the digital signal period will show if the signal could suffer attenuation. Note that inductor/capacitor combinations can cause ringing when an edge is applied, especially when the load resistance is high. A spot analysis at a defined time is also available.

    The diagram belows shows the step response of a pi filter, of 50nF capacitance per section, and 1mH inductance, to a 15V step input. The source resistance is 50 ohms, and the load resistance is 50k Ohms.

StepFilter_graph.gif (6824 bytes)

Analysis of hybrid transient protector

    High energy voltage transients, originating from lightning strikes for instance, produce two problems for single suppression devices. A silicon transient suppression diode, or metal oxide varistor switches rapidly, but may have peak voltage and total energy ratings below those produced by the transient. A gas discharge tube however has very much higher peak current capability, but also has a finite switching time. This results in a significant voltage overshoot at the load, which can be damaging.

    A so-called 'hybrid' solution, combining gas discharge tube and diode eliminates the overshoot and has high energy handling. However, a series impedance (sometimes called a co-ordinating element) is required to produce a sufficiently high voltage to allow the gas tube to switch on. The tool allows both a series resistance and inductance to be specified. An inductor allows a lower low frequency circuit impedance, but its voltage drop is dependant upon the rate of voltage increase. At high transient dV/dt, the inductor voltage drop can be sufficient to prevent the diode from switching on.

    The circuit arrangement analysed by the tool is shown below.

Hybrid GDT Circuit.gif (2900 bytes)

    In addition to specification of the basic circuit values, the gas discharge tube characteristics can be set, including the voltage overshoot against dV/dt. This data is accessed from a separate form called from the hybrid protection tool form, which is itself shown below.

Gdt-form.gif (10705 bytes)

    The voltage overshoot increases with the rate of voltage increase; this increase is logarithmic however, which means that the actual switching time decreases with decreasing risetime. A voltage overshoot figure can be set for each decade of risetime rate between 102 V/s to 1012 V/s. A typical characteristic is provided which will be adequate for many purposes.

    The tool assesses the risetime rate of any specified transient, and extrapolates between the specified points. Other characteristics which can be set include the arc voltage, which is the voltage across the gas tube in arc mode, and the holdover voltage. The latter is the voltage which appears across the gap after the current falls to a value too low to support an arc.

    Consideration of the circuit shown previously, and the nature of the switching characteristics of the gas discharge tube and the diode, shows that there are several distinct circuit conditions which may exist. For instance, the plot below from the tool shows the load voltage and gas discharge tube voltage in response to a 500 volt peak pulse, of risetime 8 microseconds and decay time 30 microseconds. The load resistance was 50 ohms, the source resistance 2 ohms, and the series elements were 5 ohms and 1mH. The diode switching voltage was 34 volts.

Gdt-graph.gif (7685 bytes)

  The plot illustrates the complex behaviour of the hybrid circuit. The blue curve shows the voltage across the gas discharge tube. The voltage rises up to the calculated overshoot value of approaching 500 volts, and then collapses rapidly to the arc voltage of 20 volts. The arc voltage is maintained until there is insufficient current to maintain the arc, after which the voltage at the gas tube rises again, and then decays.

    The load voltage (which is the same as the voltage across the suppression diode), rises initially until the diode clamping voltage is reached. After the gas tube enters arc mode, sufficient voltage is maintained at the diode due to energy stored in the inductor to keep the diode switched on for a while, after which the load voltage falls to below the arc voltage.

    When the gas tube switches off, the diode again clamps the load voltage, until there is insufficient voltage for it to conduct, when the load voltage finally decays exponentially to zero.

    If the peak transient voltage were decreased, there would eventually be insufficient voltage to switch the gas discharge tube. In this condition, the transient suppression diode would have to be capable of withstanding the pulse current for the entire transient duration.

    Designs using a hybrid device must initially consider the maximum transient voltage allowed at the load to arrive at the diode clamping voltage. D.C. or low frequency considerations will determine the maximum value of series resistance; from the protection circuit viewpoint, the bigger the value of series resistance, the better. Once the diode voltage and series resistance are estimated, the specified transient can be applied, and peak power and current ratings of the diode and gas tube arrived at using the tool.

Analysis of the thermal noise in a resistor

    Although emc is often concerned with minimising the effects of externally generated noise upon a system, it should be remembered that all resistive components produce a wide band non-periodic noise voltage, arising from thermal movement of electrons within the resistance. The open-circuit rms noise voltage produced by a resistance is given by:

            Vn = ( 4 * k * T * B * R ) 0.5 volts

where k is Boltzmann's constant ( 1.38 * 10 -23 J / oK), T is the absolute temperature (oK), B is the system noise bandwidth (Hz), and R is the resistance (ohms). The thermal noise can equally be expressed as a current generator.

Thermal Noise.gif (3649 bytes)

   The magnitude of the thermal noise is independant of whether the resistor is carbon, or wirewound etc., and depends only on temperature, the magnitude of the resistance, and the system bandwidth. Intrinsic noise within a resistor is also often called Johnson noise, after its discoverer.
    The bandwidth dependance is regardless of the actual frequency, and the noise in a 500Hz bandwidth is the same between 1kHz and 1.5kHz, as it is between 1MHz and 1.0005MHz.

Noise-form.gif (6698 bytes)

    The Thermal Noise Tool allows a spot analysis to be made, and up to 5 sets of data to be plotted, as dBuV versus resistance (log scale) plots. For constant bandwidth, the dBuVolt versus resistance graphs are straight line plots, with noise increasing with increasing frequency. A typical value for a 100kohm resistance at room temperature, in a 20kHz bandwidth system would be 15.18dBuV, or 5.74uVolts.
    In addition, the Tool calculates the maximum power that can be transferred from the resistance. This is independant of the actual resistance value, and is given by:

                Pn = k * T * B watts

    At 25oC, the available noise power is equal to 4.11 * 10 -21 watts/Hz.

    The other important consideration regarding thermal noise is the peak value of the noise signal. The calculation on the previous page provides the rms value, and not the peak value. In fact, the peak noise voltage probability density follows a Gaussian distribution, as shown below.

Noise_2.gif (4051 bytes)

    As the distribution never reaches zero, there is a finite chance of extremely high noise voltages. However as indicated in the diagram, the chance of the peak voltage exceeding 5 times the rms value is exceedingly small.

    Finally it should be noted that only components with a resistive, energy dissipating component can produce thermal noise. A capacitor for instance cannot generate thermal noise. The noise generating value of an arbitrary network of passive components is equal to the real part of the equivalent impedance of the network.


Westbay Technology Ltd, Main Street, Baycliff, Ulverston, Cumbria LA12 9RN, England

Tel: 01229 869 798    (from the United Kingdom)    Tel: +44 1229 869 798    (International)                                 Contact by e-mail