Post a Comment
Feature
Designing protective circuitry for DSL loops: Beware of pitfalls
DSL equipment requires protection from a variety of overvoltage conditions, but the need to avoid unduly degrading circuit operation complicates circuit design.
By Phillip Havens, Littelfuse LP -- EDN, 12/5/2008
|
The appropriate protective-circuit design for DSL (digital-subscriber-line) loops depends on the type of loop: Loops vary in voltage conditions and in susceptibility to attenuation and degradation in signal integrity. Therefore, protective circuitry that works well in one application may be completely inappropriate in another. A “normal” signal ranges from 2.5V in an HDSL (high-bit-rate DSL) to 260V in an DSL system (see sidebar “Normal signals”). The circuit design and choice of circuit-protection devices must take these differences into account. You should also consider the design with respect to effects of varying capacitance.
Basic protective circuitsDesigners have a number of devices to choose from—GDTs (gas-discharge tubes), thyristors, MOVs (metal-oxide varistors), and TVS (transient-voltage-suppression) diodes for overvoltage, and fuses and PTC (positive-temperature-coefficient) devices for overcurrent. The challenge is to use them effectively without unduly degrading normal circuit operation.
HDSL circuits require longitudinal protection at both the HTU-C (HDSL-transceiver-unit-central-office) and HTU-R (HDSL-transceiver-unit-remote) interfaces because of the ground connection HDSL uses with loop powering. One approach is to use a pair of TSPDs (transient-surge-protection devices) from tip to ground and ring to ground to provide overvoltage protection, preceding them with a pair of fuses—one on the tip and one on the ring—to provide overcurrent protection (Figure 1). For the transceiver side of the coupling transformer, another TSPD can provide overvoltage protection.
The ATU-C (ADSL [asymmetric-DSL]-transceiver-unit-central-office) interface and the ATU-R (ADSL-transceiver-unit-remote) interface typically use longitudinal protection; an ADSL modem at the customer premises does not, due to the absence of earth-ground connections. You may consider a GDT if the design uses a 0.5A fuse (Figure 2). A TSPD and a surge-tolerant fuse can provide metallic protection. An alternative method, with GDTs connected in the delta configuration, provides tip-to-ground, ring-to-ground (longitudinal), and tip-to-ring (metallic) protection.
Minimizing capacitive effectsOne of the drawbacks of some solid-state protective devices—TSPDs, for example—is their high, voltage-dependent capacitance, which can cause problems in three areas. Capacitance imbalance due to different biasing voltages on the tip and the ring may cause longitudinal, or common-mode, signal distortion. On-hook/off-hook/ringing transitions can cause sudden transient events to occur on the phone line, possibly changing the capacitance of the channel. Equalization established during modem training becomes suboptimal when the capacitance of the channel shifts. The capacitance of the channel changes instantaneously during the transmission of 30V-p-p signals; these nonlinear channel characteristics may lead to IM (intermodulation) distortion.
The object is to design a protective circuit that will minimize the effects of varying capacitance yet still provide protection compliant with the appropriate regulatory and standards requirements or recommendations.
|
One method is to use microcapacitance TSPDs, which can have a typical capacitance of 60 pF, or 40% less than a standard TSPD part. When 60 pF is still too much, a good alternative is to put the TSPD in series with a pair of ultrafast-switching diodes in an inverse-parallel arrangement (Figure 3). Because the diodes have capacitances of about 10 pF each, the total capacitance across the loop is approximately 15 pF. The fuse is necessary for safety and to guard against power-fault events. Although this approach works well in T3 and Ethernet applications, the high peak voltages of VDSL2 (very-high-speed DSL 2) cause the diodes to conduct sufficient current to cause linearity problems that will corrupt the coding constellations.
Figure 4a shows a circuit with two TSPDs that works well for metallic transients on CPE (customer-premises equipment); Figure 4b shows a circuit for longitudinal surges on ADSL2+. Another useful circuit is the balanced-bridge topology in Figure 4c. Yet another alternative is to apply a minimum line bias to reduce and linearize capacitance. Figure 5 shows the general topology; a 5V bias applied through two 1-MΩ resistors can serve a circuit with a POTS (plain-old-telephone-service) overlay, and a 24V bias works for a circuit without a POTS overlay.
Although there are multiple factors to consider when designing protective circuits for DSL, there are multiple useful approaches. Because of the complexities involved, designers may wish to seek advice from suppliers that offer all the circuit-protection technologies and are therefore not biased toward any one type of protective device.
| Author Information |
| Phillip Havens is telecom-sector technical manager at Littelfuse. He represents the company at many telecom-related industry meetings and helps define, direct, and support new silicon-based protection products. Havens also interfaces with customers’ technical teams to develop the most cost-effective protection for their various applications. |
| Normal Signals |
|
Though at times it’s unintentional, DSLs (digital-subscriber lines) can experience a number of voltage conditions. One important consideration is whether the loop provides POTS (plain-old-telephone service) as well as DSL. POTS involves battery and ringing voltages. Battery voltage is nominally –48V dc, or 56.6V maximum. Ringing voltage is nominally 90V at 20 Hz in the United States and 25 Hz in Europe, but it can reach 150V rms at 16 to 40 Hz. Therefore, the maximum expected voltage under normal conditions is the peak value of the maximum operating ring voltage, 150V rms, plus the maximum dc bias of the central-office battery, for a total of approximately 269V in the worst-case situation. HDSLs (high-bit-rate DSLs), on the other hand, use a 1.544-Mbps, T1-equivalent transmission rate but with half the bandwidth of a comparable T1/E1 link. The signaling levels are a maximum of ±2.5V, and loop powering for regenerators is typically less than 190V. ADSLs (asymmetric DSLs) employ transmission rates that reach 6.144 Mbps from the COT (central-office terminal) to the RT (remote terminal) and as much as 640 kbps from the RT to the COT, whereas VDSL2 (very-high-speed DSL 2) employs symmetrical data rates to 100 Mbps on short loops. Telephone engineers define overvoltages as metallic (differential mode), between the tip and the ring, or longitudinal (common mode), between the tip and ground or between the ring and ground. Overvoltages can exceed 2500V, and surge currents can reach 500A. Surges can come from nearby lightning strikes or either inductive or direct-contact ac-powerline interactions. Longitudinal overvoltages are more frequent and are attributable to power induction, power crosses, and nearby lightning strikes, provided that the line is referenced to ground. You can convert one-way longitudinal transients to metallic transients if both the tip and the ring have protective devices to ground and one device conducts before the other. |
| A glossary of overvoltage-protection devices |
|
GDTs (gas-discharge tubes): Also known as gas-plasma arresters, GDTs are glass or ceramic packages filled with an inert gas and capped on each end with an electrode. When a transient voltage exceeds the dc-breakdown rating of the device, the voltage differential causes the electrodes of the gas tube to fire, resulting in an arc that in turn ionizes the gas within the tube and provides a low-impedance path for the transient to follow. Once the transient drops below the dc-holdover voltage and current, the GDT returns to its off state. Current ratings can reach 20 kA, and capacitance ratings can be as low as 1 pF with a 0V bias. GDTs typically see use in primary protection because of their high surge rating. However, their low interference for high-frequency components makes them candidates for high-speed data links, such as in an ADSL (asymmetric-digital-subscriber-line) circuit if you use the appropriate overcurrent-protection precautions. The GDT is a poor choice for PCB (printed-circuit-board) applications and is more appropriate for primary protective for an NID (network-interface device) or an MDF (main-distribution frame). TSPDs (thyristor-surge-protective devices) better serve board-level applications. GDTs have high surge-current and low capacitance ratings. In addition, capacitance does not vary with applied voltage. MOVs (metal-oxide varistors): MOVs are two-leaded, through-hole components typically shaped like disks. Manufactured from sintered oxides and schematically equivalent to two back-to-back PN (positive-negative) junctions, MOVs shunt transients by decreasing their resistance as you apply voltage. Because MOVs’ physical dimensions determine their surge capabilities, high surge-current ratings are available. Also, because MOVs are clamping rather than crowbar devices, they can serve as transient protectors in secondary ac-powerline applications. Although MOVs are restricted from use in many telecom applications other than disposable equipment, they are useful in ac applications that require a clamping device but not tight voltage tolerances. Thyristor-surge-protective devices: A TSPD is a PNPN (positive-negative/positive-negative) device that you can think of as a thyristor device without a gate. When it reaches a switching-voltage rating, a TSPD crowbars a transient voltage, whereas the maximum repetitive-peak-off-state-voltage value ensures that it does not react to steady-state conditions. Think of it as a device that turns on with voltage and turns off with current. When the current flowing through the TSPD is less than the device’s holding current, the TSPD resets and returns to its high off-state impedance. TSPDs are appropriate for use in secondary protection. Advantages of TSPDs include fast response, stable electrical characteristics, long-term reliability, and low capacitance. Also, TSPDs are crowbar devices—that is, they act as shorts across two terminals. Thus, voltage cannot damage them, and they have extremely high surge-current ratings. However, you cannot use them directly across the ac line; you must place them behind a load. Failing to do so results in exceeding their surge-current rating, which may cause them to enter a permanent short-circuit condition. Although you may find them in other applications, TSPDs find use primarily as the principal overvoltage protector in telecom and datacom circuits. TVS (transient-voltage-suppressor) diodes: TVS diodes are constructed with PNPN junctions. During conduction, TVS diodes create a low-impedance path by varying their resistance as you apply voltage across their terminals. Once you remove the voltage, the diode turns off and returns to its high off-state impedance. Because TVS diodes are solid-state devices, they do not weaken, and their electrical parameters remain stable as long as they operate within their specified limits. TVS diodes effectively clamp fast-rising transients and are well-suited for low-voltage systems that do not require large amounts of energy to shunt. Due to their low power ratings, TVS diodes do not act as primary interface protectors across the tip and the ring; they find use as secondary protectors that are embedded within a circuit. |
| A glossary of overcurrent-protection devices |
|
Flameproof resistors: For cost-sensitive designs, 1/8 to 1/4W, flameproof metal-film resistors often serve in lieu of PTC (positive-temperature-coefficient) devices, fuses, and powerline or line-feed resistors. During a transient condition, the resistors open when the resultant energy is great enough to melt the metal the device uses. Flameproof resistors are inexpensive and plentiful, but they are not resistant to transient conditions and are susceptible to nuisance short circuits. Small resistors rarely find use as the means of protecting telecommunications equipment during power-fault conditions except in inexpensive CPE (customer-premises equipment). Fuses: Due to their stability and low series-resistance values, fuses are among the most popular options for meeting ac-power-fault requirements for telecommunications equipment. Similar to PTC devices, fuses function by reacting to the heat that excessive current flow generates. Once the fuse exceeds its interrupt rating, the center conductor opens. Fuses are available in surface-mount and through-hole packages and can withstand the applicable regulatory requirements without additional series impedance. Correctly chosen, fuses interrupt a circuit only when extreme fault conditions exist, and, when you properly coordinate them with an overvoltage protector, they offer cost-effective options for transient-immunity needs. Fuses eliminate series-line resistance, enabling longer loops; provide precise longitudinal balance, allowing better transmission quality; offer robust surge performance, which eliminates downtime due to nuisance blows; offer greater surge ratings than resettable devices, ensuring regulatory compliance; don’t degenerate over time; and come in surface-mount packaging, which uses less PCB (printed-circuit-board) real estate, eliminates the need for mixed technologies, and reduces manufacturing costs. However, fuses do not reset, so alternatives might better serve applications involving ac-strip protectors and ground-fault interrupting circuits. Positive-temperature-coefficient thermistors: During a fault condition, heat is generated at a rate equal to the interrupt rating. When this heat becomes sufficient, a PTC thermistor asymptotically increases its resistance until the device simulates an open circuit, limiting the current flow to the rest of the circuit. As the fault condition drops below the PTC thermistor’s holding current, the device begins to reset, approximating its original off-state value of impedance. Because PTC thermistors are resettable devices, they work well in a variety of industrial applications in which electrical components cannot withstand multiple low-current faults. A variety of applications use PTC thermistors. In addition to protecting telecommunications equipment, they can also prevent damage to rechargeable-battery packs, interrupt the current flow during a motor-lock condition, and limit the sneaky currents that may cause damage to a five-pin module. However, due to the devices’ off-state resistance-hysteresis curve following activation, line balance in broadband circuits may be unable to tolerate their use. Powerline and line-feed resistors: Typically manufactured with a ceramic case or substrate, powerline and line-feed resistors can sink a great deal of energy and can withstand both lightning and power-fault conditions. They are available with tight resistive tolerances, making them appropriate for applications that require precise longitudinal balance. These typically large devices, however, are not available in surface-mount configurations. Also, powerline and line-feed resistors may require either a fuse or a PTC device to act as the fusing element. You typically find powerline and line-feed resistors on line cards that use overvoltage protectors that cannot withstand the surge currents associated with applicable regulatory requirements. |
