"Ask the Experts" is moderated by Béla Lipták (http://belaliptakpe.com/), process control consultant and editor of the three-volume Instrument Engineer's Handbook (IEH). If you would like to contribute to the 5th edition by updating an existing chapter or by preparing a new chapter, or if you have questions for our team of experts, please write to email@example.com.
Q: I have just read with great interest your article "An Engineer's View of Global Warming" on Sustainable Plant (http://tinyurl.com/3rg3rz4). While agreeing with almost everything you say, I am afraid you have made a simple mistake that will give the "deniers" an easy way to refute all the good stuff you have said.
There are a couple of places you mention the "two-mile high" ice cap at the North Pole. The ice cap at the North Pole is floating, and if it melts, there will be no change in sea level. There is a "two-mile high" ice cap at the South Pole, and that sits on a large continental mass. If that ice melts, we are all in big trouble.
There is a large ice cap on Greenland, which is melting, and that will cause the sea level to rise significantly. I guess it is a polar region, but it is not the North Pole.
A: Sorry, Peter, obviously I meant Greenland and the South Pole.
On the larger issue of why does global warming cause cold and snowy winters on the East Coast, I wrote a letter to the New York Times: http://tinyurl.com/3nrx7br. If you want to read more on these topics you can refer to:
Basically, both the ocean currents and the hurricanes are gigantic heat conveyors fueled by temperature/pressure differences in the oceans and the air, which are moved by the rotation of the earth and by the Coriolis force. This force causes both the air and the water to move "downhill" as the earth's diameter reduces northward from the equator. Because of global warming, the Gulf Stream is slowing, so our winters get colder, while the hurricanes become more powerful. The behavior of hurricanes or the Gulf Stream is just as predictable as the operation of industrial heat pumps. They respond the same way to changing operating conditions (to pressure and temperature differences, etc.) and, therefore, their future behavior can be similarly predicted on the basis of their past behavior as defined by their dynamic models (relative gains, time constants, etc.)
In case of the hurricanes, at 80 °F or higher "hot spots," the water evaporates, and the moist air is pulled up from the surface of the ocean at the walls of the "eye" (this is similar to how a compressor's force pulls refrigerant vapors in the evaporator of an industrial chiller), and as the rising moist air cools (as in a condenser), rain is caused. This reduces the pressure at the "eye," causing more evaporation, faster pumping of the heat from the ocean, and higher velocity winds. The hurricane travels northwest, moved by westerly wind, and rotated by the Coriolis force (counter-clockwise). The Gulf current moves more water than all the rivers on earth, and hurricanes can generate as much force as do earthquakes.
Q: Could you explain the benefits of compensation? Why is it needed in thermocouples (CJC), two-, three- or four-wire RTDs or in thermistors (Steinhart equation)? Also, what is the difference between the two- and the four-wire transmitters? What would happen if we used 4-20 PSIG pneumatic transmitters?
A: Compensation serves to eliminate the effects of variations, such as the length of the connecting wires or the changes in ambient conditions, which have nothing to do with the variable of interest. Compensation can also serve to "linearize" the relationship between the measured variable and the output signal of the transmitter.
Whenever a transmitter or sensor operates by detecting the difference between the signal generated by the sensor and the signal generated by a reference, (such as the cold junction in case of a thermocouple), compensation eliminates the error caused by variation in the reference. Therefore, in case of a thermocouple, if the temperature of the cold junction changes due to ambient temperature variation, an error would result without compensation.
The difference between two- and four-wire transmitters has to do with the power wires. In a four-wire system, the power and signal wires are separate, making it easier to reject electrical and magnetic common-mode interference. In two-wire designs, power and signal are carried on the same pair of wires. The advantage of this design is that their sensitivity to electrical noise and loading effects are reduced.
As to the signal ranges used for the range of 0% to 100% readings of the transmitters, controllers and control valves, they do not matter much. What matters is that they accurately represent the desired value, and that they are the same for all loop components. In other words, a 4-20 PSIG or a 0.2 to 1.0 ATM range is just as good as a 4-20 mA or a 3-15 PSIG range.
The purpose of standardization is to make the components of the loop compatible, and thereby make products of different manufacturers interchangeable. One of the tasks of engineering societies, such as ISA, is to promote standardization. This is not easy (particularly not in our digital age), because standardization conflicts with the business interests of the manufacturers whose goal often is to create "captive markets" and to eliminate competition. Standardization eliminates such practices.
A: RTD sensors are typically terminated in field transmitters sending a 4-20 mA signal to a panel-mounted controller or to a DCS. Wiring between the transmitter and the RTD itself will usually be a four-wire copper cable of which only three wires are used. The RTD sensor exists as one resistor of a four-resistor bridge circuit. Two of the wires connect the RTD into the bridge. The third wire balances the bridge for the length of the copper extension wire that is used. The transmitter circuit balances the bridge adjusting one of the resistors, such that no current flows through the RTD. A fourth wire may also be used in the bridge circuit, but does not increase accuracy. Please refer to many textbooks on measurement technology listed in the ISA book catalog for more information. Béla's comments are accurate for thermocouple measurements. The various types of thermocouples are necessary to cover the range of temperatures that may be needed.
A: The benefits of compensation are concerned with linearizing the relationship between the physical process condition sensed and the output signal of a transmitter.
Linearization is useful because a PID loop has stability problems if the loop gain changes over the range (span) of the transmission signal. The PID gain term must be reduced to be stable with the highest transmitter gain. The quality of control is reduced as the non-linear transmitter goes away from the region of highest gain. Compensation also reduces errors due to offsets or zero shifts as ambient or process pressure or temperature changes.
A thermocouple has cold junction compensation (CJC) because the potential of each junction of different metals adds to (or subtracts from) the voltage of the sensing junction. The cold junction is installed to subtract from the sensing junction and is held at some reference temperature, so that only the sensing junction voltage appears at the amplifier in the transmitter. CJC may also be done by a clever arrangement of electronic parts.
The number of wires used to carry the resistance of the RTD to its transmitter depends on the resistance range of the RTD and the resistance of the wires. If the resistance of the RTD is 1,000 or more times the loop resistance of the wires, then two wires are enough. Three wires are used with 100-ohm RTD sensors. Four wires are used in a Kelvin connection for low-resistance sensors, such as the 10-ohm sensor in a motor winding.
Generally, four-wire transmitters require more power than 60 milliwatts for their operation, and a second pair of wires is used to deliver this external power to the device.
The transmission signal does not matter as long as both ends of the transmission medium know how to encode and decode the signal. The first instruments were pneumatic, partly because electricity wasn't all that reliable and partly because pneumatic power for actuators made for simpler systems. As time went on, manufacturers in the United States converged on using 3 PSI to 15 PSI ranges for the pneumatic transmission signals. When 3 to 15 PSI became standard, a manufacturer could produce only one loop component , such as transmitters or valves or controllers, and their devices could be connected with other vendors' devices to form control systems. Each vendor became free to do what it did best.
As science and technology kept advancing the bleeding edge further into society, pneumatic transmission gave way to electrical with 4-20 mA DC (after a battle over 10-50 mA DC). Again, electrical signaling was simple, but there was no power for actuators, so we still need air compressors and dryers. Microprocessors began appearing in devices for calculated compensation, leading to devices that could do control in the field, which requires digital signal transmission.
Digital transmission should be as transparent to the user as a digital telecom network, but we aren't there yet.