Editor's Note : This lengthy and insightful article is presented in four parts:
Part 1 : below
Part 2 : click here
Part 3 : to be posted December 16, 2008
Part 4 : to be posted December 18, 2008
Probably nowhere do electronic engineers face such a difficult challenge as that posed by the threats of EMI acting on linear amplifiers and sensors at the detection and pre-conditioning stages of long, sensor-instrumentation loops, such as those found in industrial-measurement and control. Obtaining and maintaining the quality of these measurements in harsh signal environments is the goal of proper signal conditioning. The use of un-amplified sensor signals, is still the most common way of conveying analog signals, representing temperature, pressure, displacement, force, and many other physical variables, to conditioning circuits via signal wires.
Controlling signal integrity in noisy, low signal level sensor applications is of paramount importance in preserving the quality of the sensor measurement, and thus is the thrust of this article. Solutions will focus on:
- signal path and forms of interference
- shielding, grounding, and ground-loops
- signal sources and measurement systems
- op amp input-filtering and EMI-hardened op amps
In this first part of the series, we will examine common ways that interference can enter the signal path in sensor-amplifier loops, the difference between differential-mode and common-mode interference, as well as demonstrate how radiated interference can be converted to conducted interference in the signal cabling. Additionally, we will demonstrate a simple way to mitigate a common type of electromagnetic interference.
Sensor/amplifier: critical path considerations
As key components of sensor/data acquisition systems, amplifiers are responsible for scaling the weak signals of the sensor (often only a few millivolts) to be compatible with an ADC's full-scale input range. One of the most common ways that annoying disturbances get injected into the signal path is through the sensor's signal communication cabling. For example, long leads act as antennas and can pick up a variety of electric and magnetic interference by coupling with environmental RF sources. The usual source of these fields is from radio transmitters, such as nearby handheld cellular phones, Bluetooth headsets, two-way radios, or wireless computer peripherals, and can easily be 3 to 10V/meter at the sensor wires. The problem arises when the cabling is long and the signal levels are low.
The offending interference can be picked up as either differential-mode (DM) or common-mode (CM) currents from stray radiated fields. The source of DM interference is from changing magnetic fields interacting with current-loops (antennas), as seen in Figure 1 . Here sensor signal lines form a current loop, within area A.
Figure 1: Cable wiring act as dipole current loops
(Click on image to enlarge)
Thus, the signal wires can be modeled as a small-loop antenna carrying RF interference currents. A small-loop is one whose dimensions are smaller than ¼ wavelength (?/4) when illuminated by a particular frequency. Therefore, from Faraday's law of induction, the RF noise induced into this loop by a magnetic field can be represented as a voltage, Vn , within that loop. Vn is proportional to the area of the loop and to the angel of incidence of the H -field to the plane of the loop, as illustrated, by the angle between them, ? . Given the magnetic flux density, B , and the direction of the field with respect to the plane of the enclosed current loop, given by the angle ? , the noise voltage at the inputs of the amplifier can be approximated with the following equation.
Common-mode (CM) interference
Unlike differential-mode interference, which is relatively easy to remedy, CM interference is a much more difficult problem. CM noise is caused primarily by electric-fields (E-field), and the effect is to raise all ground connections (cable shields, amplifier grounds, etc.) in a victim circuit above system ground potential (CM noise voltage). These CM voltages appear at the amplifier inputs and produce rectification effects, distortion and degradation of the normal signal. As we will see, in addition to using amplifiers with high common-mode rejection, maintaining balanced inputs with respect to ground, shielding of the interconnect cables, a rational grounding system, along with CM and DM filtering at the instrumentation, is the best way to maintain clean transmission of sensor signals in RFI, EMI polluted environments.
Induced noise in parallel wires
If high frequency noise sources run in close proximity to unshielded, parallel signal wires, the noise currents will usually be coupled to the victim wires through either magnetic or electric fields. These sources can be high current, industrial induction heaters, dirty power supply lines, high current switches and solenoids, 50/60 Hz high voltage power lines, etc. Further away, the coupling is principally by electromagnetic fields from two-way radios, cell phones, and TV transmitters, or radio-controlled overhead cranes.
The amount of mischief they cause depends on whether the induced noise signal is balanced or unbalanced. Figure 2 shows a situation where unshielded parallel signal wires, from a low impedance, balanced signal source (such as a thermal couple or a strain gage), run close to a source of RF noise, as might happen if the sensor wires are run in the same cable tray as a noisy power bus, or primary 50Hz or 60Hz, power cables. The fluctuating disturbances couple with the radiating wire (chiefly through mutual induction) since both the aggressor wire and signal wires run parallel to each other.
Notice that there is an unequal coupling of noise current (In ) in the aggressor wire to the signal wires. One coupled field (m1 ) is a distance d1 away from the wire a , but field m2 couples through the longer path d2 to wire b . The mutual inductances, m1 and m2 generate common-mode currents ia (cm) and ib (cm) in signal wires a and b , respectively.
That is, the induced currents travel in the same direction. This would be a far less significant problem if both currents were equal; because they would produce no noise voltage across the amplifier input impedance. But because m1 and m2 are different, ia (cm) and ib (cm) are also different. This asymmetry gives rise to a process called “common-mode to differential-mode” conversion. Thus, because m1 ? m2 , and ia(cm) ? ib(cm) , a differential current ia(cm) – ib(cm) = idm is created which produces a DM component of noise voltage (vn ) across the input impedance of the amplifier.
Figure 2: Induced noise in parallel wires. Parallel signal wires can pick up noise when the mutual inductances m1 and m2 are unequal.
(Click on image to enlarge)
The unwanted differential-mode current can be eliminated if the coupling is made equal. To cancel the inductive noise, the mutual inductances m1 and m2 must be made equal. A good way to do that is by twisting the pair of signal lines together, as close to the sensor as possible, as shown in Figure 3 . Twisting the two wires into tight loops makes the distances d1 and d2 more equal, and thus m1 and m2 are nearly the same.
Figure 3: The simplest way to reduce magnetically induced interference is to use twisted-wire pairs.
(Click on image to enlarge)
The coupling of cross-talk to one loop is cancelled by the opposite field in the adjacent loop. At least 12 loops should be wound per foot (˜37 loops/meter).
(to be continued. . . . )
About the author
Jerry Freeman is an amplifier applications engineer at National Semiconductor Corp., Santa Clara, CA. He received a BSEE from Heald College of Engineering in 1961.
Editor's note : If this article was of interest to you, also check out:
“Understanding noise optimization in sensor signal-conditioning circuits (Part 1a of 4 parts)”,
by Reza Moghimi, click here; note that Parts 1b, 2a, and 2b are linked.