Isolate the noise, not the signal
Isolators and signal converters are most often used on 4-20 mA loops, where a 24 Vdc power source provides power to a field instrument. The instrument returns a 4-20 mA signal that typically represents a measurement, such as flow, level, pressure, etc. Isolators can be introduced into the “loop” to solve simple problems such as ground loops or noise.
Some engineers think that the need for isolators and signal converters is eroding with the advent of smart instruments, isolated electronics and digital fieldbuses. However, isolators and signal converters can also be used to solve difficult or complex problems.
Isolators can be 4-wire (line- or mains-powered) transmitters or they can be 2-wire (loop-powered). A 4-wire isolator can perform signal conditioning, convert signals, split signals, boost power to the loop and many other functions. A 4-wire isolator also can be used when the isolator output must be voltage (such as 0-10 V), zero-based (such as 0-20 mA) or bipolar (such as -10 V to +10 V) (Fig. 1).
Loop-powered isolators work just like any other 2-wire DP, pressure or temperature transmitter (Fig. 2). Its output always has to be some form of 4-20 mA, but signal conversion can still be performed, as can split ranging.
An isolators that is loop-powered on the input side is a great solution when applied correctly. However, these are the most misapplied isolators. The beauty of this isolator is its simplicity. Just break the loop where convenient and insert the isolator. Wiring changes and installation costs are minimal.
Several rules must be followed if the power source in the loop is from a transmitter. The transmitter’s 4-20 mA output and its compliance voltage must power the isolator electronics and the isolator’s output. Because 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 its electronics, the isolator consumes 5.5 V from the loop. In other words, the isolator itself looks like a 275 ohm load on the transmitter. To calculate the total burden on the transmitter, add the isolator load to 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.
Input quality alarms
If a transmitter is connected to a control system, it likely has quality alarms on the 4-20 mA input to let operations know when the transmitter has failed. That is, if the control system sees a signal less than 4 mA, it knows a problem occurred. If you have to install an isolator between the transmitter and the control system, select an isolator that will not negate the input alarm strategy. This is an important consideration because not all isolator manufacturers follow the same practice.
Consider an isolator with a 4-20 mA input and a 4-20 mA output. Some isolators do not permit the output to drop below 4 mA when the input goes to 0 mA. But this is exactly what you want to avoid if you are using quality alarms.
How far down an isolator output can go depends on the type and model. Typical 4-wire isolator outputs go all the way to 0 mA because, with a 4-wire product, the output can go to 0 mA. When an isolator is loop-powered, the output cannot go to 0 mA because there is no power to run the isolator electronics. If the isolator is analog, the output can go to 3.0 mA before power loss. Microprocessor-based isolators are more power-hungry, and their outputs cannot go below 3.8 mA.
If the DCS is an older system that has a quality alarm with a fixed setpoint of 3.2 mA, use an analog isolator to allow the quality alarms to function.
It is often very convenient to power 2-wire transmitters from an isolator. Typically, such an isolator has three terminals: TX, +IN and -IN. The TX power supply is current-limited to 25 mA. This is ample for 2-wire transmitters %%MDASSML%% even those that have a diagnostic failure, which causes the current to drive to 24 mA.
We have seen cases where some microprocessor-based transmitters are labeled as loop-powered, but during their startup cycle they draw 35 mA. In cases such as this, the TX output will not suffer permanent damage, but it also will not provide sufficient current to start the transmitter. Fortunately, this is rare.
It is quite common to share process signals between two different systems, such as two control systems; an emergency shutdown system (ESD) and a control system; one DCS and a data acquisition system; and so on. Generally, it is unacceptable to create one series loop between the transmitter and two systems. You do not want a series loop because if you have to disconnect the input at one system for maintenance purposes, then both systems would lose the signal.
One solution to sharing a variable with two systems is to use a single isolator (Fig. 3). One system is declared the primary system, and it powers the transmitter. In Figure 3, the primary system is the ESD. The isolator isolates the primary loop from the secondary loop. The DCS is the secondary loop. Maintenance can disconnect the input to the DCS without affecting the signal going to the ESD.
The architecture in Figure 3 is very common, but there is a weakness: if you have to disconnect the input to the ESD, the DCS also loses the signal. Disconnecting the ESD input removes power from the transmitter. The alternate solution is to use a “splitter.”
A splitter is a 4-wire isolator with one input and two outputs (Fig. 4). All I/O and power are isolated from each other, and the splitter powers the transmitter. The difference between the area isolation and splitter approaches is that you can disconnect either control system for maintenance without affecting the signal going to the other system. This is a very popular solution in applications such as custody transfer and isolating validated systems from non-validated systems in the BioPharm market.
Loops rarely start out overburdened. But over time devices get added to the point where the drive capacity of the transmitter has been exceeded. Nominally, a loop-powered transmitter powered with 24 V will drive into 600 ohms. If your receiving devices are still using 250 ohm input impedances, it does not take many to overburden the loop. An isolator solves the problem by providing a convenient way to add more power to the loops.
Bucking power supplies
Some DCS manufacturers offer lower cost 4-20 mA input cards, but the tradeoff is that this card must power all the loops. There is no problem when all your inputs are from loop-powered transmitters. But when you have 4-wire magmeters or other mains-powered transmitters, both sides of the loop are trying to source the 4-20 mA and the result is too little or no current. A simple loop-powered isolator solves this problem (Fig. 5).
Microprocessor-based isolators offer some solutions that are not found in analog isolators, such as custom linearization. Products often have 85- or 128-point linearizers that can very simply address applications such as linearizing tank level to volume in a non-linear tank, square root extraction, signal limiting, characterizing pH to reagent demand and valve linearization.
Noisy process signals
In ideal control environments, input signals to isolators would be clean and flat, with no extraneous process or plant noise. Such environments don’t exist, unless your plant is in a lab. There will always be unwanted noise accompanying your desired control or monitoring signal. Often these fluctuations are part of the process, and searching for ways to soften or eradicate them could sometimes be detrimental to the process.
In other instances, recently installed equipment, grounding problems or unknown entities can conduct noise onto your input signals. With any of these situations, a digital or microprocessor-based isolator can quickly and easily remedy the situation.
Signal isolators, splitters and converters have come a long way since their days of simply protecting instruments from ground loops. These simple and inexpensive devices can often solve complex and expensive problems.
|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. For the past three years, he has specialized in signal conditioning products including isolators, signal converters, alarm trips, remote I/O and related devices.|
The straight skinny on signal conditioning
Signal conditioners are the nucleus of any test measurement system. They receive output from the transducer, carry out basic processing of the signal and pass the signal on to the display, storage and analysis components. Signal conditioners may also supply an excitation voltage to power the transducer.
In a typical measurement system, signal conditioning equipment is located between the transducer and the display. It conditions the transducer signal into a form the display can use. Some transducers have the signal conditioning electronics built into the body. This type of transducer can provide a high-level, low-impedance output signal that is immune to noise contamination caused by radio frequency interference or electromagnetic interference. The disadvantages include a limited temperature environment, resulting from the electronic components and the large size of the transducer.
Other transducers have some or all of the signal conditioning built into an inline unit located near the transducer but away from any extreme test environment. This is a compromise with advantages in some test situations. Disadvantages are the limitations in the number and adjustability of the inline unit’s signal conditioning capabilities. The unit may also be located with limited accessibility.
The most common situation is to have the transducer mounted directly on or close to the test object. Cable runs can carry the transducer’s output signal to the signal conditioning equipment.
You can use certain types of signal conditioners with some transducers that share a common operating principle. Use a strain gauge signal conditioner with a strain gauge pressure transducer, a strain gauge load cell or with strain gauges to determine stresses in materials. Use a charge amplifier with piezoelectric accelerometers or piezoelectric pressure transducers. The test measurement professional must still ensure the signal conditioner is compatible with the selected transducer and the signal conditioner’s specifications meet the test measurement requirements.
For some transducers, the cable also carries the excitation voltage the transducer requires and the signal conditioner provides. The main advantage of this type of system is its flexibility. Typical bench-mounted or rack-mounted signal conditioners seeing use in this kind of system have adjustments you can tailor to suit the transducer, display and test requirements.
The most common electrical output from test measurement signal conditioning equipment is a dc voltage range linearly proportional to the transducer’s input range. A typical dc voltage range is 0 to 10 V, where 0 V is the output from the signal conditioner when the transducer input is zero, and 10 volts is the output when the transducer input is at full scale.
Regardless of the types of sensors or transducers you use, the proper signal conditioning equipment can improve the quality and performance of your system.
An excerpt from the book, Fundamentals of Test Measurement Instrumentation , by Keith Cheatle, published by ISA Press.