Now that all of these factors in the second part of our article on false alarms have been presented, what are the next steps?
First there is a need for explanation of some of the problems we face in the electrical environment in which we live.
Power line noise is, and always has been, a natural by-product of making electricity, its distribution and its use. If the demand for power increases beyond the capacity of the company's generator in one area, or if heavy industry overloads the system, the corrective action taken by the electric company can cause spikes and transients to appear on the power line. This process is known as "notching" or "grid switching". The electric companies try to carry it off smoothly, but it usually results in a "notch," which is a brief dropout in power that is followed by a brief surge as normal power is restored.
There is also the "power faction connection." This is a process of switching capacitor banks in and out to compensate for heavy industrial loads. It also causes a brief burst of electrical noise.
Lightning strikes are not the most common source of power-line noise, but they are certainly the most devastating. Even when the strike is many miles away, surges and spikes measuring thousands of volts may show up at electrical receptacles. Mark Twain once remarked that, "Thunder got all of the credit for the noise, when it was actually the lightning that did all of the work." Well, as dramatic as lightning is, and it draws more than its share of attention when it comes to power line disturbances, on a day-to-day basis, far more electrical noise is created right in our buildings and homes by noisy electrical loads.
Fluorescent lights, copiers, electric typewriters, heating, ventilating and air conditioning, coffee makers, power tools, arc welders, vending machines, and other like tools and appliances generate noise that goes back into the electrical system as they operate. All electrical loads make some noise, at least when they are turned on and off. The noisiest are those that have high peak current demands -- those that have inductive components, such as electric motors or solenoids, especially those that use solid state switching devices, such as those found in thermostats and light dimmers. And don't forget, computerized systems with their peripherals are frequently a part of the problem, not just victims of it. Computer monitors, disk drives, and printers, for example, frequently have large start-up current requirements that can cause transients to travel to and from the peripherals.
Why then are today's systems so much more sensitive than they used to be?
Well, during the past 10 years most computer manufacturers have changed from using linear DC power supplies to modern switch-mode or switching power supplies. The switching power supplies are lighter, smaller, more efficient, and insensitive to changes in power line voltage. However -- and its always the "however" that gets you -- since they operate by switching on and off rapidly and drawing a lot of current during each cycle, they can generate a lot of noise if they are not connected to a very low-impedance power source. Inadequate wiring or some other high-impedance power source is usually the problem here. The other big problem is chips themselves are more vulnerable to noise than they used to be. Why? In each new generation we pack more and more transistors into the same microscopic space. And as the individual transistors get smaller and smaller, the amount of electrical overstress they can survive gets lower and lower. Modern semiconductor devices can now be disrupted by as little as half a volt of noise. Over 10 volts of noise starts to destroy them.
What's the bottom line? Power line noise enters a computer from two sources-normal mode and common mode. Normal mode noise is a noise voltage potential between line and neutral. Common mode noise is a noise voltage potential between line and neutral and ground. Computers are better protected against normal mode noise than they are against common mode noise.