April 10, 1997

Solve transient-protection problems with devices and topology

William Russell, Semtech Corp

CMOS-transceiver ICs are exposed to harsh transient conditions that lightning strikes, ESD, and transmission power-line faults cause. The right combination of devices and topologies can provide the secondary protection you need in T1 line cards.

Voltage transients are abnormally high pulses of voltage that exceed the circuit’s normal operating voltage. These transients generally are random in nature and may occur for tens of nanoseconds to a few milliseconds. Three sources of transients in telecommunication lines are the natural phenomena of lightning, ESD, and ac-power faults.

Electronic systems that directly connect to external data and I/O lines are especially susceptible, and T1 line cards are no exception. The delicate structure of the CMOS devices makes them susceptible to damage from voltage and current surges. This vulnerability requires that you protect these devices by external transient-suppression circuitry.

Lightning is the most common cause of transients in electronic systems. At any time, there are approximately 1800 thunderstorms in progress around the world, generating 100 lightning strikes/sec on the earth’s surface. Mean peak current for the first stroke is 20 kA, with subsequent strokes decreasing by 50% or more. Although a direct lightning strike is dramatic, lightning also produces intense electric and magnetic fields that can couple into nearby power lines, communication lines, and circuit wiring, causing catastrophic or latent damage to semiconductor equipment. The magnitude of the induced voltage varies with the distance from the strike. The complex web of telephone-network-communication links makes them especially susceptible to lightning-induced-transient damage.

You can generally define lightning impulses by combination waves—one wave for voltage and another for current. The waves are defined as double-exponential pulses with a specified rise time and duration, or decay time. For example, an 8/20-µsec impulse-current waveform has an 8-µsec rise time and a 20-µsec decay time to 50% of the peak.

Another cause of failure in T1 lines is ESD, which results from a sudden and violent redistribution of electrons between two objects. The amplitude of the static discharge can exceed 30 kV, with currents reaching 30A or more, and the discharge lasts less than 1 nsec. The human body is one of the most common ESD generators. It capacitively stores charge with respect to ground. The voltage potential with respect to earth ground can exceed several thousand volts. It is common, in fact, for a person to develop voltages as high as 15 kV; voltages as high as 35 kV are also possible, but rare.

Although the ESD pulse contains little energy, its fast rise time and high power can devastate sensitive semiconductors. Failure may result from junction shorting, oxide puncture, or melting of the device metallization.

An ac-power fault, or power-cross fault, results from contact between the T1 line and an ac-power line. This situation generally occurs because of the proximity of telephone lines to power-distribution lines. The magnitude of ac voltage resulting from power-cross faults can exceed 600V ac. Although T1 lines are more susceptible to power-cross faults, electromagnetic coupling from nearby ac-power lines can induce potentially harmful voltages on any telecommunication lines.

The best strategy for dealing with lightning- and ESD-induced transient events is to divert transient current away from the T1 transceiver IC using transient-voltage-suppression devices that are shunt-connected across the protected line. During a transient event, the transient current is shunted through the device, resulting in a lower transient voltage across the protected circuit.

Surge-suppression devices that are shunt-connected can be either clamping or crowbar devices. Manufacturers have optimized each type for a certain transient condition. Under normal operating conditions, the clamping device looks like a high-impedance path to the protected circuit. Ideally, the device appears as an open circuit, although a small amount of leakage current is present. When a transient voltage exceeds the line’s normal operating voltage, the device becomes a low-impedance path. The transient power dissipates within the device and is limited by the maximum allowable junction temperature. Clamping devices automatically return to a high-impedance state when the line voltage returns to a normal level.

A crowbar device begins to break down with a positive resistance until the device reaches a break-over voltage, at which point the device "snaps" back to a low on-state voltage. The low on-state voltage means that the device dissipates less power and thus provides a higher surge-current-handling capability than does a clamping device. The disadvantage of crowbar devices is that the current through the device must fall below a vendor-specified holding current for the device to return to a nonconducting state.

Available technologies for parallel-protection elements include gas-discharge/surge arresters (commonly known as "gas tubes"), metal- oxide varistors (MOVs), and solid-state devices (transient-voltage-suppressor (TVS) diodes and TVS thyristors). Each type of device serves a specific application. In some cases, you may combine two or more device types to use the advantages of each.

Gas tubes employ an internal inert gas that ionizes and conducts during a transient event. The gas is contained in a glass or ceramic envelope with specialized electrodes placed at each end. When the voltage across the terminals reaches a certain level, the gas ionizes, causing the device to spark over, or "fire." At this point, the gas tube becomes a low-impedance path for the passing transient. The circuit voltage must fall below the gas tube’s holdover voltage before the device extinguishes and returns to a nonconducting state. Because the internal gas requires time to ionize, gas tubes can take several microseconds to turn on. In fact, the reaction time and firing voltage depend upon the slope of the transient front.

Gas-surge arresters have a finite life span. As the gas tube wears out, leakage currents and firing voltages increase until total failure occurs. Also, if an impulse overstresses the gas tube, the hermetic seal may be breached, allowing the internal gas to escape and render the device useless. The gas tube’s high-current-handling capability allows you to use the tube as a primary surge protector at connections to the outside world. The gas tube’s high variable-firing voltages make the tube unsuitable for protecting solid-state circuitry at the board level. Break-over voltages typically range from 90V to 1 kV.

MOVs are nonlinear devices whose resistances vary with applied voltage. MOVs consist of a ceramiclike material, usually in a disk shape. You achieve high-transient capability by increasing the disk’s size. The disk’s surface is coated with a highly conductive metal to ensure uniform current distribution and is encapsulated with a plastic-epoxy material. The interface between the zinc-oxide and the metal-oxide matrix material is roughly equivalent to two back-to-back pn junctions. You can, therefore, view the MOV as several pn junctions in a series-and-parallel configuration. This configuration gives the MOV its characteristic VI curve, similar to back-to-back zener diodes.

MOVs gradually degrade with each transient event. Granular interfaces overheat and begin to short, resulting in a gradual decrease in breakdown voltage. Eventually, the MOV fails, and the device achieves a permanent low-impedance state, thus resulting in a loss of protection. MOVs turn on in a few nanoseconds and have high clamping voltages, ranging from approximately 30V to as much as 1.5 kV. The MOVs’ high parasitic capacitance makes them unsuitable for use on digital T1 lines.

TVS thyristors are solid-state devices constructed with four alternating layers of p- and n-type material. The resulting structure is similar to an SCR whose gate is controlled by an avalanche zener diode. TVS thyristors are crowbar devices that switch to a low on-state voltage when triggered. Because the on-state voltage is low, the TVS thyristor can conduct high surge currents. Again, the current through the device must fall below the holding current before the device returns to a nonconducting state. TVS thyristors do not wear out or exhibit the large overshoot voltage of gas tubes. Additionally, TVS thyristors respond in nanoseconds and have operating voltages starting at approximately 28V. The devices’ unique characteristics make them well-suited for many telecommunication applications.

TVS diodes are solid-state pn-junction devices. A TVS-diode junction employs a large cross-sectional area so that the diode can conduct high transient current. By controlling such factors as junction depth, doping concentration, and substrate resistivity, the manufacturer can vary the device’s target voltage. TVS diodes are clamping devices. When the transient voltage exceeds the circuit’s normal operating voltage, the TVS diode becomes a low-impedance path for the transient current. The device returns to a high-impedance state after the transient threat passes. TVS diodes do not wear out and have no degradation of the electrical parameters, as long as you operate the device within vendor-specified limits.

A primary attribute of a TVS diode is its reaction time. Avalanche breakdown theoretically occurs in picoseconds. However, this breakdown is difficult to measure, so manufacturers specify TVS diodes to respond almost instantaneously. TVS diodes are available in a range of operating voltages. Traditional discrete-device voltages range from 5 to 440V; recent innovations in TVS technology have yielded devices that operate at 2.8 and 3.3V.

TVS diodes have an inherent voltage-dependent capacitance, which can cause excessive signal degradation in high-speed data applications. Companies manufacture low-capacitance devices by placing a low-capacitance rectifier diode in series, but opposite in polarity, with the TVS diode. This placement adds another capacitor in series with the junction capacitor of the TVS diode. The resulting total capacitance is less than the smallest component in the series, according to the well-known formula for capacitors in series. By carefully choosing the rectifier, you can reduce the effective capacitance by approximately two orders of magnitude. TVS diodes’ fast response time and low clamping voltages make these diodes ideal for use as board-level protectors for semiconductors and other sensitive components.

Additional protection devices that you use in T1 applications include positive-temperature-coefficient (PTC) resistors and Schottky diodes. PTC resistors are overcurrent-protection devices that you use as series-protection elements, and they limit current during ac-power crosses in telecommunication applications. Under fault conditions, the PTC heats, and its resistance sharply increases, limiting the fault current to the protected device. After the fault condition is removed, the PTC cools and resets, allowing the protected circuit to function normally. Often, the devices exhibit increased resistance after you reset them.

Manufacturers form Schottky diodes by creating a junction between a metal surface and a silicon substrate. The resulting device has unique electrical characteristics that set it apart from a conventional pn-junction rectifier. These characteristics include a low forward-voltage drop and fast switching time due to the absence of minority carriers. You can use Schottky diodes to clamp CMOS-IC lines to the supply rail, thus preventing the transceivers’ internal protection diodes from turning on.

Finish with the appropriate circuit

You need a combination of protection devices to provide effective transient protection for T1 interfaces (Figure 1). For outside lines, install the primary protectors at the network interface. However, these protection devices, usually spark gaps or gas tubes, do not suppress transient voltages and currents to levels that are safe for line-card ICs.

Add secondary protection elements at the board level to suppress the let-through energy from the primary protectors, as well as from induced lightning and ESD transients. The secondary protection elements must dissipate high transient energy and provide minimum loading capacitance. A device with too little power-handling capability is destroyed during equipment testing, and too much capacitance causes the digital waveform to fall outside T1 pulse template specifications.

It’s a good idea to implement secondary transient protection at the power-supply- and signal-line interfaces. Good design practice normally requires signal-line-protection elements on the line and IC sides of the line transformers. Line-side elements absorb the bulk of the transient energy, thus protecting the transceiver IC and the transformers. Chip-side protection guards the IC transceiver against fast transient energy that is coupled through the transformer because of winding capacitance. The type of line protection you require depends on the lines to which your device connects.

The outside-line-protection circuit in Figure 2 meets the common-mode (line-to-ground) and differential-mode (line-to-line) transient-immunity and power-cross requirements of Bellcore 1089, FCC Part 68, and UL 1459. The design also meets the ESD-immunity requirements of IEC 1000-4-2. The protection scheme is a cost-effective, multiple-stage design that protects the transformer and the line IC using readily available components and minimizing required board area.

The first stage of the design employs a PTC resistor to limit current during ac-power-cross faults, meeting the requirements of Bellcore 1089 and UL 1459. PTC resistors are available in pretrip resistance values of 4 to 7 ohms. The PTC resistor also limits current to the protection elements under transient conditions. It slightly reduces the transmitted signal’s amplitude; transceiver manufacturers often specify adding a small-value resistor on the IC side of the transmit line to reduce this effect. Line-feed resistors that combine a fuse with a resistive element are sometimes substituted for the PTC device. The disadvantage of line-feed devices is that you must replace them after the fault condition is removed. Also, changes in device resistance after resetting can cause the lines to become unbalanced. For this reason, designers often use a combination of surge-rated resistors and fuses instead of PTC resistors.

The second stage of the design contains the main elements for meeting the lightning-surge requirements of Bellcore 1089 and FCC Part 68. During the metallic-lightning surges, a low-capacitance TVS diode (LC01-6) provides voltage and current limiting. The TVS diode’s fast reaction and low clamping ability allow the diode to shunt the surge current with a maximum clamping voltage less than the transformer-saturation point, which is typically less than 50V. During longitudinal (common-mode) surges, you use a solid-state crowbar device comprising Z₁ and Z₂ as the suppression element. Because of the line transformers’ high potential-isolation rating, the common-mode suppression element does not require a precise clamping threshold. You can choose a device with a break-over voltage of approximately 275V, vendor-rated to handle 100A. Note that you can eliminate Z₁ and Z₂ as long as your system voltage does not exceed the high potential-isolation ratings of the T₁ line transformers and power-supply system.

Implement the final stage of protection on the transceiver side of the transformer to protect the IC from any residual energy that is coupled through the transformer. On the receive line, use resistors to limit the amount of current that enters the IC. The values may be large and without signal attenuation, because R_TIP and R_RING are high-impedance inputs. On the transmitter-output pins, Schottky diodes provide clamping for residual transients, because large series resistors attenuate the digital signal beyond template specifications. A high-voltage capacitor prevents dc flow into the transmitter transformer. On the power-supply line, the TVS diode clamps power-supply transients and capacitors provide high- and low-frequency power-supply decoupling.

Don’t forget inside-line protection

The approach you use to protect intrabuilding metallic lines is similar to the outside-line-protection scheme, except that you can use fewer, and less expensive, devices (Figure 3). This design meets the intrabuilding lightning-surge and 120V-ac-power-cross requirements of Bellcore TR-NWT-001089. The protection elements also provide ESD protection to IEC 1000-4-2 Level 4.

You use PTC resistors in the first stage of the design to provide the intrabuilding ac-power-cross protection that customer-premises equipment requires. You can substitute a time-delay fuse and a 1W, surge-rated series resistor for the PTC resistors in some designs.

The second stage of the design provides the main elements for meeting Bellcore 1089’s 100A, 2/10-µsec intrabuilding lightning-surge requirement. During the metallic-lightning surge, the low-capacitance TVS diode shunts the surge current with a maximum clamping voltage that is less than the transformer-saturation point. You use an enhanced diode-bridge configuration to provide the low capacitance of the TVS-diode protection circuit. You use the bridge rectifiers as steering diodes to safely route the incoming surge through the TVS diode. You can implement this configuration using discrete devices, as long as you choose components to provide the surge and low-capacitance characteristics you desire.

However, implementation with discrete components can be tricky. First, selecting the right bridge diodes is important. The devices you select must have high forward-surge capability and fast switching to be effective against ESD- and lightning-induced surges. Second, device layout is critical to reduce overshoot voltages associated with trace inductance. Finally, high component count can cause design problems when pc-board space is at a premium. Alternatively, new devices are available that combine the surge-rated diode bridge with a high-power TVS diode in one SO-8 package. Common-mode surge protection is unnecessary if the line transformers are rated for a minimum isolation voltage of 1500V.

The recommended protection scheme for the transceiver side of the transformer is nearly identical to the one you use on the outside-line-protection circuit—combining resistors, capacitors, and fast-switching diodes to protect against residual transient energy. A surge-rated, low-capacitance diode array is an alternative to Schottky devices to clamp residual surges to the rail (Figure 3). You use a TVS diode and decoupling capacitors to protect the transceiver power supply.

References

1. Stearman, Greg, Larry Stillings, and Roger Taylor, "Board-level secondary protection provides added reliability for T1 digital line cards, Parts 1 & 2," Compliance Engineering, Spring 1993.
2. Curtis, Jon, "Meeting Bellcore’s TR-NWT-001089 surge requirements," Compliance Engineering, Summer 1993.
3. Level One Corp Application Note, "LXT360/361 line protection circuitry," January 1996.
4. Stringfellow, Dr Michael F, "Lightning," Power Quality & Assurance, September/October 1995.
5. Russell, William, "Defuse the threat of ESD damage," Electronic Design, March 6, 1995.
6. Semtech Corp Application Note #SI96-16, "Outside line lightning protection for T1/E1 CMOS IC transceivers," 1996.
7. Semtech Corp Application Note #SI96-17, "Intra-building lightning protection for T1/E1 CMOS IC transceivers," 1996.

| EDN Access | Feedback | Table of Contents |