Process control instrumentation: Isolator and signal converter basics
Greg Feliks, Moore Industries -- Plant Engineering, 12/12/2007 11:56:00 AM
The need for isolation has flourished since the 1960s when the 4-20 mA current signal started becoming the accepted standard for field transmitters to send measured process variables to control room receivers such as indicators, recorders or control devices. The problem arose when signal wires had paths to ground through the field transmitter at one end of the signal loop and the receiving device at the other end. When the voltage (or ground potential) at the two ground points was not equal, an unwanted current “i” was formed (Fig. 1). The closed current path was formed by the copper wires used for the 4-20 mA signal and the ground – thus this current “i” became known as a ground loop. Therefore, the receiving device actually measures 4-20 mA ± i. That ground loop error was unacceptable and a way to eliminate it was needed. |
| Fig. 1. The ground loop “i” is formed when there are two grounds at different potentials. |
The ground loop forms when three conditions are true: there are two grounds; the grounds are at different potentials; and there is a galvanic path between the grounds. To eliminate the ground loop, we need to eliminate any one of these three conditions. We cannot always control the number of grounds, and it is often not possible to just lift a ground. The ground may be required for the safe operation of the electronic device, or the ground exists because the instrument is in physical contact with the process, which is in physical contact with the ground. From a practical standpoint, we cannot reach into the earth and regulate the voltage at these ground points. We can, however, use some technology to “break” the galvanic path between the two grounds – that is exactly what an isolator does. A current cannot form when we break the conductive path between the differential voltages. We have eliminated the ground loop.
In today’s marketplace, we are sometimes fortunate enough to have transmitters with isolated outputs and receiving devices with isolated inputs; the manufacturers of these products have embedded an isolator on the input or output card. Often though, we have to insert an isolator in the signal loop between the transmitter and the receiving device (Fig. 2). This may be because the built in isolation is insufficient. Proper basic application of the isolator requires us to first understand the characteristics of an isolator and how the isolator gets its power to function.
 |
| Fig. 2. An isolator breaks the galvanic path between two grounds. |
Transformer isolation
Again, the first and foremost duty an isolator must perform is to find a way to break the galvanic path between the sides of circuits that are tied or “grounded” to different potentials. A galvanic path is defined as a path in which there is a direct electrical connection between two or more electrical circuits that allows current to flow. Breaking this galvanic path can be accomplished by any number of means including electromagnetic, optics, capacitive, inductive and even acoustics. With most industrial measuring equipment, the two prevalent methods chosen for galvanic isolation are transformer and optical. Because this article pertains to the process control industries, only these two types will be discussed.
 |
| Fig. 3. The transformer provides isolation between the input and output circuits. |
Transformer isolation – often referred to as electromagnetic isolation – uses a transformer to electromagnetically couple the desired signal across an air gap or non-conductive isolation gap (Fig. 3). The electromagnetic field intensity is proportional to the input signal applied to the transformer. Transformers are very efficient and fast at transferring ac signals. Since many process control signals are dc, they must be electrically “chopped” into an ac signal so they can pass across the transformer. Once passed, they must be rectified and amplified back into the desired dc signal output.
 |
| Fig. 4. Galvanic isolation is obtained using an opto-coupler circuit. |
Optical isolation uses light as a means of transferring a signal between elements of a circuit (Fig. 4). The opto-coupler or opto-isolator is usually self-contained in a small compact module that can be easily mounted on a circuit board. An optical isolation circuit is comprised of two basic parts: a light source, usually an LED acting as the transmitter, and a photo-sensitive detector, usually a phototransistor, acting as the receiver. The output signal of the opto-coupler is proportional to the light intensity of the source. The insulating air gap between the LED and the phototransistor serves as the galvanic separation between the circuits, thus providing the desired isolation between two circuits at different potentials. Optical isolation has better common-mode noise rejection, is usually seen in digital circuits, is not frequency sensitive, is smaller and can sometimes provide higher levels of isolation than transformer isolation.
Two-way vs. three-way isolation
Two common terms used within the process control industry with respect to isolation are two-way and three-way isolation. It is important to understand that all isolators are not created equally. Often you will find isolation specifications that detail what the isolation levels are from input to output. This is often referring to two-way isolation, and is suitable for loop-powered isolators.
However, many manufacturers fail to mention or outline their isolation details when their isolators are line- or mains-powered – that is, when the isolator requires 24 Vdc, 120 Vac or 240 Vac to operate its circuits. In these instances, ensure that you have an isolator that has full three-way isolation. Three-way isolation is defined as input-to-output, power-to-input and power-to-output isolation.
If the isolator is powered by a dc supply, many manufacturers use common signal wires between the output and the power input. In these situations, it is very important that you guarantee that the potential level of your receiving device and your power supply are at the same levels – something that is very hard to dictate. Otherwise, you could have problems with common mode noise, or unwanted output signal errors created by a failing switching power supply.
EMI/RFI filtering
The effects of Radio Frequency Interference (RFI) and Electromagnetic Interference (EMI) can cause unpredictable and non-repeatable degradation in instrument performance, accuracy and may even lead to complete instrument malfunction or failure. When these unwanted types of noise find their way into your measurement circuits, it often results in off-spec product, reduced process efficiency, plant shutdowns and sometimes dangerous safety hazards.
Some of the more common sources of interference are mobile and stationary radios and handhelds, static discharge, large solenoids or relays, ac and dc motors and welders. The three basic approaches that are commonly used to protect an electronic instrument from the harmful effects of radio frequency and electromagnetic interference are a well made case or housing, robust board level design and layout, and terminal strip protection.
Because RFI can “squeeze” through even the smallest cracks, an isolator’s DIN-style housing should be designed so there are no openings, uncovered mounting holes or exposed PCB areas. These housings should use solid aluminum cases, which help block stray RFI and EMI – something that plastic housings cannot do.
Other design features that further enhance the ability to negate the effects of noise include low pass filters, terminal strip common ground plane to the low side of the filters, strategically-placed capacitors and inductance filters (LC) and the use of full spectrum ceramic RFI/EMI filters. In fact, there are some isolators available that have filtering that guarantees less than 0.1% error at radiated noise levels of 50 V/m between 20 MHz and 1,000 MHz.
Specially designed circuitry offers the last level of defense. Advanced PC board layouts in good isolators often use multi-layer designs with ground planes, “lossy” ferrite beads and strategically-positioned capacitors to minimize the effects of unwanted noise that could lead to instrument inaccuracies or failures.
Common- and normal-mode rejection
Just as RFI and EMI can adversely affect control signals, so can normal- and common-mode noise. This noise can often be at frequencies close to line frequencies (50/60 Hz), or at much higher frequencies. Normal-mode noise is most often found in ac circuits where there is an unwanted signal or “noise” that exists between the neutral and hot lines. Common-mode noise is most often found between the neutral and ground wires. However, there can also be common-mode noise between the hot and ground wires. While normal-mode noise can be present in dc measuring circuits, common-mode noise is the most prevalent, and causes the most incorrect readings or measurements.
Isolators and dc monitoring circuits must be designed to tolerate a certain level of common-mode noise. A circuit’s ability to reject this noise is typically expressed in a form similar to 120 dB at 50/60 Hz. It’s usually the job of isolation amplifiers to provide this level of protection. Figure 5 depicts where common-mode voltage (CMV) would be present in a dc measuring circuit. Rejection circuitry built into good isolators will allow an instrument to accept high levels of common- and normal-mode noise without adversely affecting the desired signal.
 |
| Fig. 5. This is where common mode noise would be present in a dc measuring circuit. |
Ambient temperature effects
When choosing instrumentation, one of the many considerations is the amount of heat the electronics will be exposed to over its installation lifetime. While field instruments often possess sufficient ambient temperature operating specifications, many DIN-mounted instrument suppliers assume that their equipment will be installed in a climate-controlled environment.
You will often see DIN rail instruments with a maximum ambient operating temperature rating of 60 C. Properly sized metal cases allow adequate ventilation and act as heat sinks to assist with heat dissipation. Plastic cases and housings are less effective heat sinks. While heat is the major culprit of failure of electronic components, colder climates can also present myriad challenges. Some isolators are suitable for installation in areas with temperatures as low as -40 C, and as high as 85 C.
Isolator power
Isolators can be 4-wire (24 Vdc, 120 Vac or 240 Vac powered) transmitters or they can be 2-wire (loop-powered) either on their output side or their input side. Why we would select one power configuration over another relates to the application being addressed.
The 4-wire isolator is used any time the isolator output has to be voltage (example: 0-10 V) or zero-based (example: 0-20 mA) or bipolar (example: -10 V to +10 V). A 4-wire isolator usually sources its current output and generally the drive capacity is around 1,000 to 1,200 ohms (Fig. 6).
 |
| Fig. 6. This is an example of a 120 Vac powered 4-wire isolator. |
Some specialty products will drive into 1,400 or 1,800 ohms. A 4-wire isolator can have plenty of power to drive custom circuits, so these isolators will be used for custom signal conversion, like 0-10 V to 0-200 µA. It can power dual isolator or splitter configurations and some offer an internal 24 Vdc supply to power 2-wire transmitters on the input to the isolator. Also available are 4-wire isolators that have a current sinking output that looks like a loop powered transmitter to the receiving device.
Isolators that are loop-powered on their output side are powered just like any other 2-wire differential pressure, pressure, or temperature transmitter and this tends to be the lowest cost installation (Fig. 7). The output always has to be some form of non zero based current, typically 4-20 mA, but signal conversion can still be performed as can split ranging. When powered with 24 Vdc, these isolators typically drive into 600 ohms.
 |
| Fig. 7. Isolator loop-powered on the output. |
The third power configuration is a great solution when applied correctly, but it is also the most misapplied isolator. Isolators that are loop-powered on their input side have several rules that must be followed. The beauty of this isolator is its simplicity. In figure 8, pretend that the isolator did not originally exist and after startup it was discovered that isolation was required. You just break your loop where convenient and insert the isolator. Wiring changes and installation costs are minimal.
 |
| Fig. 8. Isolator loop-powered on the input. |
Note that the only power source in the loop is from the 4-wire transmitter in the field. The transmitter’s 4-20 mA output must power the isolator electronics and the isolator’s output. Because available power is limited, the isolator output load is limited to 250 ohms. The receiver’s input impedance can be anywhere from 0-250 ohms and it should be a fixed load. Additionally, there can be no voltage on the output of the isolator. To run the isolator electronics, the isolator consumes 5.5 V from the loop or, in other terms, the isolator itself looks like a 275 ohm load on the transmitter. To calculate the total burden on the transmitter you have to add the isolator output load to the 275. The total load could then be as high as 525 ohms plus wire resistance. That is not usually a challenge for a 4-wire transmitter but it can be for a loop-powered transmitter limited to 600 ohms.
Compliance voltage
Some users use the term compliance voltage. When referring to isolators, compliance voltage is the internal voltage source driving the output current. Compliance voltage is related to drive capacity (in ohms) through Ohm’s Law. Moore Industries generally expresses isolator outputs in terms of drive capacity, such as “4-20 mA into 1,000 ohms.” From this spec, to determine the compliance voltage, the equation is:
V = 20 mA x 1,000 ohms = 20 V
A valve positioner may need 11 V to operate. The spec refers to the compliance voltage behind the 4-20 mA control signal to the positioner. To express that in terms of resistance, the equation is:
R = 11 V ÷ 20 mA = 550 ohms
So the positioner is a 550 ohm load on the 4-20 mA signal. If you add 50 ohms to cover wire and termination resistance you are up to 600 ohms. You probably do not want to use a loop-powered isolator in an application that has only 600 ohms of drive capacity.
Isolators are useful devices to solve instrumentation problems in process control applications. However, be sure to check the specs carefully. Not all isolators are created equal.
Isolators can be used for many applications in process control beyond just eliminating ground loops and conditioning signals. Part 2 of this discussion will present a selection of advanced applications for isolators and signal converters.
Greg Feliks is a senior application engineer at Moore Industries. He has been in the industrial and process industries for more than 20 years involved with field instrumentation, recorders, controllers, control systems and various other types of industrial controls. Feliks owned a manufacturer’s representative firm in the Atlanta area, and worked for Dynisco Instruments and Eurotherm before joining Moore Industries. He has held positions in sales, marketing and product management. For the past three years, he has specialized in signal conditioning products including isolators, signal converters, alarm trips, remote I/O and related devices.
There are no other articles related to this article.Talkback
Related Content
Related Content
Sponsored Links
