A couple of weeks ago, while out in the field, someone asked me how one went about hooking a 4-20 ma current loop up to a data logger. Since I happen to have done just that recently for a project up near Seattle, including building up the power supply panels that are a big part of the solution (the picture below is the field deployed panel with logger monitoring a pneumatic signal to a control valve), I thought I would do a string of blog posts on the topic.
The DC power supply panel is the gray box. The data logger is inside it. The transmitter that is converting the pneumatic signal to the control valve from pressure to a 4-20 milliamp signal is the black square thing tie-wrapped to the railing in the upper left hand corner of the picture. The picture below is a close up of it and the actuator it is monitoring.
Current loops are a very common way to transmit data from a sensor to a remote location and a common standard is to provide a linear signal that varies from 4-20 milliamps as the measured variable shifts over the field sensor’s range. Other standards exist, like 10-50 milliamps, but they are not common and usually limited to process control applications.
Current loops offer advantages in terms of noise immunity and in terms of allowing information from a sensitive, but low gain measuring element to be accurately transmitted over long distances. For instance, a 100 ohm platinum RTD is a very accurate way to
measure temperature. But the resistance change associated with a temperature change is very modest; fractions of an ohm per degree F.
RTD measurements are typically made by applying the RTD in a resistance bridge and using the change in voltage as in indication of the change in temperature. Since the changes in resistance are small, the associated changes in voltage are small, typically on the order of millivolts per degree F.
Frequently, in the field, the induced voltages from the conductors serving our machinery can exceed these voltage levels by several orders of magnitude. I discovered this early in my career when troubleshooting a chiller interlock circuit. I kept picking up 10 – 15 vac in a circuit that, as near as I could tell, was isolated from any power supply. Then I realized that my control conductors were running in a long cable tray in parallel with the large conductors carrying hundreds of amps to a different chiller in the plant. In essence, the long run of parallel wires was acting like a transformer, inducing a voltage in the interlock circuit from the power feeders. Turning off the chiller made the induced voltage to away.
As you can imagine, 10-15 volts of induced signal would totally obscure a millivolt signal on a cable carrying information from an RTD to a controller. Shielding helps address this issue by blocking and channeling away the undesired signal, but is also a bit tricky. For instance, if you ground the shield wire (either intentionally or accidentally) at both ends of the cable, then the shield becomes a current carrying conductor. This is because we think of ground as being zero volts, but if you were to actually measure the “ground” voltage relative to a common point in the building, you would discover that there are minor voltage differences.
So, if you connect a conductor (for instance a shield on a cable) to “ground” at two different points, by virtue of fundamental physics, a current will flow due to the voltage difference across the conductor. And, again by virtue of physics, flowing current generates electro-magnetic fields which can (you guessed it, by virtue of physics) couple to adjacent conductors and induce voltages and currents. Incidentally, if you want to learn more about shielding, I learned a lot from a book titled Grounding and Shielding Techniques in Instrumentation by Ralph Morrison.
Even if shielding were perfectly implemented and eliminated the potential for noise in the measured signal, the wires themselves that are carrying the millivolt signal from the RTD to the controller can cause a problem. Specifically, wire, even a good conductor like copper, has a resistance. So, a long run of wire adds to the resistance of the RTD, and with out special compensation circuits, can be interpreted as part of the signal. The Control Design Guide contains an example of how much this can impact the information from an RTD in a typical HVAC application.
This was another lesson from the field that I learned early in my career when, after measuring and re-measuring, I discovered that the lead resistance in a 1,000 ohm copper RTD circuit, which amounted to 1-1/2°F, was by far, greater than the temperature change the control system needed to detect to make the decision to start or stop a chiller (about 1/2°F).
Compounding the problem was the fact that resistance of most conductors varies with temperature. So, in my little troubleshooting situation, I also discovered that when the
mechanical rooms were cool the lead resistance error was different from when the mechanical rooms where hot. Again, the change was more than the change I was trying to measure because in the winter the mechanical rooms could be in the low 60’s°F while in the summer temperatures at the ceiling where the wiring ran could exceed 120°F.
Adding a 4-20 milliamp transmitter to the RTD circuit can address some of these issues. For one thing, since the signal is a current rather than a voltage, it is fairly immune to the impact of induced voltages. In addition, as long as the power supply is adequate, the wires can be run for literally miles serving multiple controllers and indicators with out degrading the signal. Of course, the transmitter itself will introduce its own errors, but usually (but not always) the cost is worth the benefit.
In addition to the technical reasons discussed above, current loops are attractive because they are a common standard supported by just about every manufacturer. That means that if you can pick up a signal from a 4-20 milliamp current loop, then you can measure
just about anything from temperature to flow to pressure to toxic gas levels.
So, there you have it; an overview of why you might want to use a 4-20 milliamp current loop. Next we will take a look at how they work and then we will look at how one goes about applying one in the field as an input to a typical data logger.
David Sellers
Senior Engineer – Facility Dynamics Engineering
A Field Perspective on Engineering 