Achieve EMC-compatibility for industrial RS485 networks

Preserve RS485 signal integrity: Electromagnetic compatibility (EMC) transient protection requires matching the performance of protection components to the characteristic of transceiver component. The suggested circuits can help a designer reduce risk of project slippage due to EMC problems. See diagrams, tables.

By Thomas Kugelstadt July 17, 2013

Modern industrial networks rely on robust RS485 communication links for long distance data transmission, so protecting signal clarity can be a challenge. These networks must operate in harsh industrial environment and hence are subject to strong electromagnetic interference in the form of large transient voltages caused by electrostatic discharge, electrical fast transients, and lightning strikes. Data rates for long distance communication range from 10 kbps up to 250 kbps over distances of up to 4000 ft (1200 m) and lately even up to 6000 ft (2000 m). Due to its differential signaling technique across twisted-pair cable and the ability to operate reliably in a high-common-mode environment, RS485 has become the interface workhorse in industrial applications.

To prevent costly network downtimes, due to transceiver damage caused by high voltage and current transients, the International Electrotechnical Commission (IEC) has developed transient immunity tests for electrostatic discharge (ESD), electrical fast transients (EFT), and surge transients. For proper network implementation, understand each of these transients, their application, and design solutions to prevent component damage.

Transient comparison

The ESD test simulates the electrostatic discharge of a human onto electronic equipment. A test pulse has a rise time of approximately 1 ns and lasts less than 100 ns. A test sequence consists of 10 positive and 10 negative pulses with a one-second pause interval between each pulse (Figure 1). ESD has the lowest energy content of all transients.

The burst test simulates switching transients caused by inductive switching, relay contact bounce, and so on. This test applies a sequence of test pulses called burst. A test pulse has a 5 ns rise time and lasts about 400 ns. A burst consists of 75 pulses applied at a repetition rate of 5 kHz, followed by a 300 ms pause. A test sequence comprises six 10-second bursts with 10-second pause intervals and produces 14,000 pulses per minute. An EFT pulse train has about 300 times more energy than an ESD pulse of the same test voltage.

The surge test simulates switching transients caused by lightning and switching heavy inductive loads. The test distinguishes between an open-circuit and a short-circuit pulse-shape with different rise times and pulse durations. During application, the test pulses are often referred to as a combination waveform. Surge transients are 1000 times longer than ESD or burst transients. Additionally, the low source impedance of the surge generator assures high surge currents at high voltages. A test sequence consists of five positive and five negative surge pulses with a one-minute pause interval between pulses. A surge transient has approximately 100 times more energy than an EFT pulse train, and about 30,000 times more energy than an ESD pulse of the same test voltage.

Legacy protection methods

Figure 2 shows two classic protection schemes intending to prevent bus transceiver damage and bit errors from high-voltage transients and other common-mode noise.

The circuit in Figure 2a uses steering diodes that clamp the signal lines to VCC (common collector voltage) or ground potential in case of a transient event. This circuit stems from a legacy bus termination scheme using Schottky diodes to minimize reflections of an otherwise unterminated bus. In the case of signal reflections, which typically rise to levels above VCC and below ground, these diodes limit the reflected signal to the maximum levels of VCC + VFW (full wave voltage) for positive reflections, and 0 V ground (GND) – VFW for negative reflections.

There are two major issues with this performance. The first is that the circuit only works in small common-mode environments due to the early clamping action of the diodes. The EIA485 standard, however, requires solid data transmission across a common-mode voltage range from -7 V to +12 V, thus making this circuit RS485 noncompliant.

The second issue is that every time these diodes conduct or clamp, which literally occurs at every signal transition, huge peaks of diode currents are diverted or steered either toward VCC or ground. These current peaks contribute to large radiated emissions from the circuit into the environment, so are most likely to fail any electromagnetic interference (EMI) compliance test.

While the circuit in Figure 2b uses a more complex protection scheme, the components applied might prevent it from suppressing fast transients. For example, low-cost transient voltage suppressors (TVS) of the SMAJ, SMBJ, and SMCJ type possess large junction capacitances and, therefore, long response times. In the event of fast transients, voltage overshoots of up to 120 V can occur before the TVS diodes start clamping, making transceiver damage inevitable.

The resistors with positive temperature coefficient (PTCs) increase their resistance in the presence of high currents. Their response time lies in the range of milliseconds, which is far too slow for surge transients let alone ESD and EFT pulses.

Common-mode chokes often are the preferred tool to filter common-mode noise. Their application, however, requires detailed study of the choke characteristics and how they fit into an application. It is not feasible to just copy a protection circuit from an audio or USB-3.0 reference design and expect it to work in a long-distance RS485 application.

Plenty of literature suggests adding filter capacitors to steepen the filter response. However, care must be taken to prevent the filter transfer function from creating an unnecessarily high peak around the cut-off frequency. Fast transients, such as ESD and EFT, have a wide bandwidth exhibiting frequency components from 3 MHz to 3 GHz, some of which could be unintentionally amplified by the filter peak.

Moreover, filter capacitors should match in value. Large component tolerances cause different cut-off frequencies between the signal lines, thus converting common-mode into differential noise and leading to data errors.

Modern circuit protection

Figure 3 presents a variety of protection circuits for various levels of transient protection.

Circuit in Figure 3a uses a fast, low-capacitance (75 pF), 400-watt TVS with breakdown voltages of 13.5 V and -7.5 V, which assures compliance with the required common-mode voltage range of 12 V to -7 V specified in EIA485. Series resistors, Rs, in the A and B signal lines are among often-overlooked necessities. These resistors provide current limiting during a transient event, as well as create the necessary voltage drop to keep the TVS turned on for the duration of the transient.

The circuits in Figure 3b aim for higher surge protection levels and, therefore, also require use of a transient blocking unit (TBU) and a thyristor transient suppressor (TISP).

The TBU is current and voltage triggered. During a surge event the current through the TBU rises to the current limit level, in approximately 10 ns. At this point, an internal voltage switch disconnects the load (TVS and transceiver) within approximately 1 μs. During the remainder of the surge, the TBU device remains high-impedance with leakage currents of less than 1 mA in the protected state of very low current and voltage at the load.

The TISP device is a symmetrical, voltage-triggered, bidirectional thyristor. Overvoltages are initially clipped by breakdown clamping until the voltage rises to the breakover level, which causes the device to crowbar into a low-voltage on-state condition. This low-voltage on state causes the current resulting from the overvoltage to be safely diverted through the device. The device switches off when the diverted current falls below the holding current value.

As in the previous case with the TVS and series resistors protecting the transceiver SCR, for increased surge transient levels, the TBU and the TISP protect the TVS and its following circuitry.

For even higher surge levels, the circuit in Figure 3c replaces the TISPs with a gas discharge tube. Gas discharge tubes (GDTs) are designed to prevent damage from transient disturbances by acting as a “crowbar” to create a virtual short-to-ground circuit during conduction. When an electrical surge exceeds the defined breakdown voltage level of the GDT, the gas becomes ionized and rapid conduction takes place. When the surge passes and the system voltage returns to normal levels, the GDT returns to its high-impedance (off) state.

Table 1 lists the protection levels for the circuits in Figure 3, while Table 2 presents the bill of materials.

Table 1: Protection levels for circuits in Figure 3

Courtesy: TI

Table 2: Bill of materials for circuits in figure 3

Courtesy: TI

Matching characteristics

The key to successful EMC transient protection is to match the performance of the protection components to the characteristic of the transceiver component. While the suggested circuits cannot replace the due diligence required at the system level, they help the designer to reduce the risk of project slippage due to EMC problems.

– Thomas Kugelstadt is a senior systems engineer with Texas Instruments. Edited by Mark T. Hoske, content manager, CFE Media, Control Engineering, mhoske@cfemedia.com.

ONLINE

At www.controleng.com/archive August, see references, more information.

www.ti.com/esd-ca 

www.ti.com/industrial-ca 

www.ti.com/interface-ca 

Key concepts

  • Industrial network signal integrity requires protection against transients
  • RS485 serial communications, with distances up to 6000 ft, can be susceptible to interference
  • Ensure transient protection matches network as implemented

Consider this

Network signal reliability is needed for reliability signal delivery. Will this tutorial help?