Electrical noise can be reduced or eliminated by using chokes, shielded isolation transformers, dedicated circuits, and shielding.
In protecting our sensitive electronic equipment from electrical noise, is there a set of priorities we can follow? The answer is YES, thanks to the work of the writers and teachers of the industry's power quality consensus standards: FIPS PUB 94 and the IEEE Emerald Book. Our being alerted to noise-related events has given us guidelines to solve many of the problems.
Noise on the phase conductors
For those situations where electrical noise is traveling on our phase conductors, we know we'll need a device in series with the circuit. This device may be as simple as a coil or choke, which offers a "kickback" to high-speed switching surges while letting the power frequency pass through. Or, the device may have to provide other functions, such as voltage stabilization or continuous energy supply to the load in question.
The former device, which is an inexpensive fix for high-speed noise on wires, is similar to a balun (a small choke coil used in the telecommunications industry to protect signal circuits).
The latter device fits into the area of power conditioning tools such as voltage regulators, line conditioners, motor generators, and UPS systems.
As you can see, we have a great range of choices. However, we also have a wide window of costs. A note of caution here: Because noise on phase conductors may not be your main problem, don't focus too much attention here. You may miss other more important areas of noise rejection.
Noise circulating in the ground
Over the years, we've come to respect this type of noise far more than we ever expected. We knew there was electrical noise circulating in our grounds, but we never saw the effect of such voltage levels until they overpowered the signal levels of our digital logic systems. (Remember, these signal levels have gotten smaller and smaller with advanced systems.) When our digital signal drivers and receivers began to send shutdown messages while trying to operate on false noise voltages, we rightly increased our concerns about this kind of noise. We began investigating how to stop the noise from disrupting and/or damaging our systems.
Remember, we're dealing with a highspeed burst of noise, or a high-frequency impulse usually seen between a neutral wire and the equipment case ground. This phenomenon is proof that we have multiple ground references for our sensitive system.
As with noise on phase conductors, a balun can be used here, but as a common-mode choke slowing down or softening the impact of the circulating noise. We can also use a shielded isolation transformer in our power circuit; this device will turn around the high-frequency burst that is traveling on the phase conductors and looking for ground potential. In effect, an shielded isolation transformer redirects common-mode noise back into the noncritical areas of our building's wiring. This device is a recommended, modest cost complement to any power circuit supplying digital power supplies.
Transient voltage surge suppression
For noise traveling on metallic pathways, we shouldn't forget transient protection methods used for power and signal conductors, particularly when these metallic conductors enter and leave areas of transient voltage danger, such as the earth. Here, our protective tool is designed to shunt the damaging noise burst into the ground and away from circuits for sensitive electronic equipment.
By using a series of protectors (some at higher voltages; some at lower voltages), this protection scheme, along with the inductive capacity of our facility, will attenuate and reduce the noise transient to a non-disruptive level.
What about induced noise?
While we're able to insert devices into our conductors and wiring system to counter noise traveling on our phase and/or ground conductors, what do we do about induced noise? This is electrical noise that couples itself onto circuits that are physically near to each other.
The most obvious answer is that we have to impose some protective material between the noise and the devices requiring protection. In other words, we have to provide a shield of some sort. This, in fact, is the generic answer to mitigating the effects of induced noise.
First, let's consider the simple case of insulated wires near each other. As shown in Fig. 1 (on page 31), bursts of noise are coupled through the air from one wire to an adjacent one. The noise travels between the two wires by means of an electric field, which is established by the flow of current in one wire (noted by arrow along wire). It's coupled into the second wire by the electric field that's circling both wires (noted by curved arrows). Remember your basic electricity and the right hand rule: Current in the direction of the thumb and a field of force in the direction of the fingers.
Whatever the force effect that may arrive at the second wire, it has to come from the noise effect of the first wire. This is the same as was detailed in our example of Part 2, where we had the primary and secondary wiring of a noise-rejecting transformer running in the same through. The noise in the primary wire was coupled onto the secondary wiring.
The recommended way to help cancel this induced noise effect is to use a dedicated circuit, as shown in Fig. 2. In this single-phase example, we have one hot wire, one neutral wire, and one or two ground wires, all serving one load and all enclosed in one metallic conduit. With this installation, we've accomplished two things. First, we've kept individual circuits apart from each other. This is especially important if the circuits are serving dynamic loads that are subject to high-speed pulses or sharp changes in operation.
Second, we've enclosed each circuit in its own shield, if you will, namely the metallic conduit. It will deflect and absorb the field energy.
Shielding of a wider scope
Shielding can take the form of metallic conduit, as noted above, or it can take the form of a large screen that covers all six surfaces of a room that's full of sensitive equipment. This screen can be shielded wire or even what is called "mu" metal to protect a workstation or terminal. Whatever the form used, the principle is the same: Block the effects of air-transported disturbances and redirect conducted disturbances.
Based on prior experience, almost all data and signal circuit providers and users adamantly stress the importance of shielding. However, many installations are calling for unshielded construction where shielding has been the norm. Our advice is: Be careful when removing this excellent protection. Let's cite a case history to emphasize the point.
A downtown high-rise building located in a major city was having problems with its unshielded telephone handsets. We first tried to identify and eliminate the electrical noise on the handset wire (the cord from the handset to the wall outlet) and were unsuccessful.
We then tried to identify the frequency of the noise and its possible origin. The noise was found to be in the FM broadcast range and was coming into the building on one of its sides, and only on certain floors.
Obviously, shielding the entire building would be too costly a solution. Thus, the phone manufacturer was contacted for possible suggestions for shielding the individual handsets. The solution was a special shield design for each handset and a special plug attachment, all at a nominal expense.