Figure 1. Side View of a Typical TVS Package

Abstract: Transient Voltage Suppressor, TVS, devices are often used for protection from voltage transients in electronic circuits. Sometimes they are used for purposes, such as ESD protection, for which they are not characterized by the manufacturer of the part. Specifications of such devices and applicability to other uses are discussed. Potential problem areas are highlighted.

Discussion:  Transient Voltage suppressors are widely used to clamp transients due to lightning surges or other causes on signal and power lines in equipment. One particular part, the SMAJ48, is a silicon avalanche diode that is popular for protecting 48 volt power circuits in equipment. Inspection of a data sheet from one of the many manufacturers of this component shows the maximum current (typically 40 Amperes) is specified for an 8.3 millisecond single half sine wave on top of the rated load. The peak pulse power of 400 Watts is specified for the 10 microsecond (rising edge) by 1000 microsecond (falling edge decay) lightning surge waveform. The maximum peak current of 4.3 Amperes at rated clamping voltage for this part is also rated for the 10x1000 microsecond waveform. These specifications indicate to me the SMAJ48 is designed primarily for relatively slow waveforms. The literature claims response times of less than 1 picosecond, but no data is provided for such a fast waveform.

Table 2 shows some additional characteristics from a manufacturer's data sheet. Note that even though breakdown occurs at about 70 Volts, the clamping level rises to over 90 Volts at its rated current of 4.3 Amperes.


This part is sometimes used for protection from very fast transients, such as ESD. Several problems can result when the SMAJ48 is used on such transients, including unexpected results. First, ESD currents are often much higher than four Amperes, so just from device resistance alone one would expect the clamping level to be significantly higher for many ESD events. Worst than that, package and connection inductance is much more important for the high di/dt that many ESD events exhibit. For an ESD example, let's take a current change of ten Amperes in one nanosecond and assume five nH of package and connection inductance (including PWB traces). Five nanohenries is not a lot of inductance. Compare the results of the ESD example with 4.3 Amps in ten microseconds implied from the specification sheet.

• 10 us waveform from data sheet:  E = Ldi/dt = 5nH * 4.3Amps/10us = 2.15 mV! ( a very small voltage)

• 1 ns rise ESD current: E = Ldi/dt = 5nH * 10Amps/1ns = 50 Volts!
• This is a very significant voltage to add to the clamp voltage of  >94 Volts that occurs at 4.3 Amperes.
• The total voltage drop might be on the order of 200 Volts if we neglect device capacitance.
  

The inductance of the package is not even specified on the data sheet I have!

In order to use this part for ESD protection, one would have to build a test jig to measure its characteristics. If this is not done, one cannot count on reliable, repeatable protection from ESD currents using this part.

Another parameter from Table 1 of interest is the device capacitance. For ESD, device capacitance may affect the result as much as the clamping action itself. The device capacitance varies with voltage and can go from about 100 pF at about 48 Volts bias to about 1000 pF at lower voltages. Five nanohenries of inductance and 100 pF of capacitance makes a nice tuned circuit at about 225 MHz lowering to about 70 MHz at 1000 pF. This could cause an interesting oscillatory response in some instances. There may be an impact on EMC emissions performance as well from the addition of this tuned circuit.

Summary: Protection devices should be characterized for the intended use. If the manufacture's specification does not cover the planned use, the user should characterize the part. It may be necessary to do it separately for each manufacturer to be used. The ESD example given shows how parasitic inductance can have a significant effect.