Ground loops: Causes and cures
Analog wiring from a sensor to I/O point is prone to certain types of problems than can disrupt accurate readings. Analog wiring is still very common, with 80+% of plants saying they have 4-20 mA communication with instrumentation. Engineers that have been trained to troubleshoot digital networks may not know how to handle these systems, so we offer this tutorial as a help for this older but still critical technology. Ground loops are particularly troublesome since they are capable of serious signal disruption and their effect can be intermittent.
For a ground loop to occur two things must be present:
At least two different grounds that are at different potentials.
These grounds must have a galvanic path or circuit established between them. Usually the circuit is completed when the process signal wire is connected from the transmitter (ground potential #1) to the receiving device (ground potential #2).
Ground loops cause problems by adding or subtracting small current or voltage levels to or from the process signal. The I/O point receiving the signal can't differentiate between the desired and corrupted signal, so the readout or DCS will not reflect true process conditions. (Ground loops are also common in audio equipment, including home sound systems, where they turn up as annoying hums.)
There are two solutions:
Eliminate all but one ground.
Isolate each ground from the others.
Eliminating grounds may not be possible for some instruments (e.g., thermocouples) because they require a local ground for accuracy, or they must be grounded for personnel safety.
Figure 1: Two-wire isolator application.
When multiple grounds cannot be eliminated, the solution is to use signal isolators. These devices break the galvanic path (the dc continuity) between grounds that are at different earth potentials. Without this path, there is no way for any stray current or voltage to reach the receiving device. Moreover, an isolator also eliminates another problem: ac continuity noise, otherwise known as common mode voltage.
There are two basic types of signal isolators:
Four-wire (line- or mains-powered)
Two-wire isolators (Figure 1) draw power from the 4-20 mA process signal loop, so they require no additional power supply or lines which can save on wiring costs. Most two-wire isolators are output loop powered (OLP), but there are two-wire isolators that are input-loop powered (ILP) too. Output loop powered isolators are more convenient because most control system current input cards offer optional power for two-wire transmitters. This makes wiring an OLP isolator to a two-wire transmitter very simple.
Figure 2: Four-wire isolator application.
When evaluating isolator specs, look at the input impedance or burden to the loop. Many devices have input impedances that are around 250 ohms. At 20 mA, this can represent a full 5 volt drop on the loop. If the loop power supply cannot support this burden, the loop will become weak or overloaded. Some isolators have an input impedance as low as 2 ohms which helps maintain voltage.
A four-wire isolator (Figure 2) is powered by an external source, such as 117 V ac or 24 V dc. Four-wire isolators work well in applications where power is readily available.
What else can an isolator do?
Isolators can perform several tasks, besides eliminating ground loops:
Signal conversion : An isolator can perform signal conversions, such as from 1-5 V to 4-20 mA, to allow devices with incompatible signal types to interface with one another.
Divert and protect signals : Using a signal isolator, you can send the output from one transmitter to a second location, protect expensive monitoring/control equipment by eliminating common electrical paths, or create a buffer between devices to allow interruption of one leg of a loop without impacting the other.
Amplify (boost) signals : If a signal is weak, or if additional instruments need to be installed on a loop, a signal isolator can boost the signal.
Solve "bucking power supplies" : When two devices (such as a 4-20 mA transmitter and a DCS) are trying to source power to a loop, the result is a non-functioning loop. When neither device can be eliminated, the solution is a signal isolator. The isolator can operate with powered inputs from both sides, thus restoring normal operations on the loop.
This article was prepared with information and graphics supplied by Moore Industries-International.