IT'S A DIRTY WORLD out there-and I'm not talking about urban air pollution. I'm referring to electro magnetic noise, ranging in frequency from slightly more than DC all the way to visible light. Although some audio signals have received a lot of research attention, others have received very little notice. In Part I of this article, I will cover power line related noise. Next time, I will discuss radio-frequency interference (RFI).
In many cases, noise alters the desired signal so subtly that you might never realize that you are hearing something totally foreign to audio. Only in the worst cases does noise become so obvious that you must take remedial measures.
Power line related noise is seldom heard. Most decent audio designs do not pick up enough of this noise to cause an audible problem. On the other hand, some types of power line noise can have a dramatic effect on reliability.
Most people assume that the power provided by their local utility resembles the 60Hz output of their signal generator, except with much higher power behind it. It does not.
(European readers should insert 50Hz whenever I refer to the US standard 60Hz.) If you were to connect an oscilloscope to the power line (I do not recommend doing this unless you really know what you are doing), you would be amazed by what you would see. Instead of a clean 60Hz at exactly 117V AC (RMS) or about 165V peak, you would see all sorts of distortions, spikes, glitches and noise, depending on where you were and what time of day it was. You would also notice that the voltage varies by more than 10 percent and probably is not even close to “exactly 117V AC. ” That's the kind of problem this article addresses.
A Summary of Nasties
Ever since the computer started taking over our lives, researchers have been looking for ways to feed the number-crunchers properly. Not until that time did many otherwise knowledgeable people realize how “dirty ” a power line could be.
Although the names of the different kinds of “dirt ” change, this is a summary of what you might find.
Transients. This is actually a broad term that refers to noise of relatively short duration (up to about 100msec).
You can break this category down in to two types-voltage spikes and oscillatory decaying transients.
Voltage spikes are very short pulses that result from large devices switching on and off the power grid.
If the device switches at the “zero voltage ” point of the sine wave, no spike is generated. If it switches at the maximum or minimum peak, however, a very large spike is generated. This pulse can often exceed several thousand volts.
If there is considerable distributed inductance and capacitance in the power line, the pulse will excite a resonant circuit composed of that inductance and capacitance, and it will “ring. ” This will spread out the energy in time, and instead of one huge pulse, you will have an oscillation that will decay in a few cycles.
While the peak voltage will be less, the duration of the event will be longer.
Mother Nature also contributes to this category with her own form of electrical noise-lightning. If you happen to be in the way of her at tempt to neutralize a charged cloud, you could find yourself hundreds of thousands of volts “above ground. ” About the only thing worse than that is the EMP (electromagnetic pulse) caused by an atomic explosion. Even if you are not in the immediate area of a lightning strike, the power lines can carry voltage surges of thousands of volts for miles.
Hash. This refers to erratic noise caused by a poor contact. It usually occurs at the peak of the waveform as the voltage becomes high enough to arc across the contacts. Two very common causes of hash are electric blanket controls and aquarium heaters. Generally, the cause is some thing in your home or at one of your immediate neighbors. If you live at the ocean, you will see light shows caused by the power line transformer insulators arcing over. All it takes is a heavy fog, and the breakdown volt age across the insulator will become less than the distribution voltage.
This will also cause hash.
Voltage Flicker. This refers to short (only a few cycles) changes in the amplitude of the power line voltage.
It generally occurs in industrial areas where heavy users intermittently draw large amounts of current. For example, a large spot welder can cause flickers.
Sags and Surges. While the utility power supply is normally considered to have a zero source impedance, it has enough source impedance to contain voltage variations caused by load changes. You might think this problem exists only in industrial areas, but my refrigerator caused enough of a voltage sag to make my computer do strange things. If you have ever noticed the lights dim when you turn on your power amplifier, you have just caused a voltage sag in your house.
A surge occurs when a major load goes off line and the utility regulation takes a few cycles to readjust itself. When this happens, the power line momentarily goes above the “average ” value.
on what is feeding your part of the world and what is using the power, you can have significant harmonic distortion. This can cause higher than normal supply voltages and higher than normal current through supply filter capacitors.
Brownouts, Outages and Failures.
A brownout is a deliberate attempt by the utility to lower total power consumption by lowering the power line voltage. This will happen only during peak load times or when a failure in the grid feeding your area has occurred.
An outage is a short (typically one or two cycles) interruption of your power resulting from a fault in the distribution system. If the fault is far enough away from where you are, the system will recycle itself, and your power will come back on. Watch out for transients when this happens.
A failure is (obviously) when nothing comes out of the outlet. This can happen when there is a fault near you or over an entire area when the entire grid fails or when you don't pay your electric bill.
More Bad News
In what is referred to as a “landmark ” study of power line disturbances, re searchers determined that a transient disturbance would happen an aver age of about 114 times per month and other disturbances would happen an average of about 15 times per month.
These disturbances were defined as being large enough to cause program errors, memory loss and system damage in computers. IBM funded the survey ten years ago, and I think the problem has gotten worse since en.
My experience with power line problems indicates that they probably account for many otherwise inexplicable failures. This is not limited to computers, but includes all sorts of equipment, including audio equipment in my home and in broadcast and other professional uses.
What To Do
If you have taken everything to this point seriously, you might be tempted to pull all the plugs out of the walls and start buying dry batteries in case lots. Actually, this is not a bad idea.
The telephone company learned a long time ago that the most reliable power source is a room (literally) filled with lead-acid batteries. In fact, until about the 1940s, most broadcast facilities derived their A and B sup plies from motor generators and batteries. Because you cannot control what happens outside your home and very few of us have the luxury of generating our own power, the only option is to tame the problem just before it gets into your equipment.
You can isolate sensitive equipment from the power line right in the power transformer. Most power transformers are designed to have relatively poor frequency response because they are meant to operate at essentially one low frequency. In most cases, the magnetic coupling becomes very inefficient at high frequencies and thus limits most forms of noise.
However, if the transformer is constructed with the primary and secondary in very close physical proximity (to minimize cost and losses), it contains sufficient capacitance be tween them to couple high-frequency phenomena from the utility line into your equipment. One way of lowering the capacitance between the windings is to separate the primary and secondary coils. There is a practical limit to this, as the farther you separate them, the greater the transformer losses become. A better way is to add a grounded single turn of conductive foil between the primary and secondary windings. This is called a Faraday Shield and is commonly found in good power trans formers, although it is difficult to find as an off-the-shelf item.
The entire Stancor catalog contains only one series of step-down transformers that include electro static shields. The PCA series is designed to provide plus and minus regulated 18V DC. The transformers have an electrostatic shield around the primary, which is internally grounded to the transformer frame.
There is still enough stray capacitance to conduct some noise around the shield, so if you really want to eliminate noise in the secondary, you can shield both the primary and the secondary in what is called a “box shield. ” This means that you would enclose the primary and secondary windings in separate conductive “boxes. ” If you do this properly, you will have less than 0.005 uF between the primary and secondary.
This is called a high-isolation transformer and is readily available, although only as a 1:1 transformer.
The Topaz Series 30 Ultra Isolator is a box with a line cord as the primary connection and a duplex receptacle as the secondary connection. You inert this device between the power line and your sensitive equipment. A 1kVA transformer costs about $600, but if you are really serious about your equipment, it is money well spent.
Unfortunately, this transformer only isolates and does not regulate.
Your equipment will still be subject to sags, surges and brownouts. Topaz and some other manufacturers do offer high-isolation transformers that also regulate. I use a Topaz Line2 Power Conditioner for my computer.
A 1kVA device like mine costs only slightly more than a high-isolation transformer of equivalent rating (about $700). This transformer solved my computer's problem. It no longer develops memory lapses when the refrigerator starts.
Another transformer design lends itself to a combined regulation /isolation application. Ferro-resonant transformers are specially designed to operate in a magnetically saturated mode. Because of this, the output changes little over a wide range of in put voltages. If you include proper shielding of the primary and secondary, it will also operate as a high-isolation transformer. Two disadvantages of this transformer are that it is relatively inefficient (about 80 per cent) and it generally hums or buzzes loudly enough to be annoying.
One characteristic of the low efficiency is that the 20 percent of power lost is a continuous function. In other words, a transformer rated to produce 800VA on the output would require about 1kVA on the input at maximum load. A constant 200VA would be converted to heat and lost, even with no current being drawn from the secondary.
One other disadvantage of ferro resonant regulators is that current limiting is “built into ” their operation. Once you exceed the maximum designed current, it will operate in a constant-current mode. While this might be useful if you want short circuit protection, it can cause lower than expected power-supply voltages because the regulating transformer might limit the peak current draw.
I do not recommend using a ferro resonant regulating transformer in series with a class B transistor amplifier because the amp's peak current requirement is much higher than the average current requirement. You would have to use a much larger transformer than you might expect, which would aggravate the efficiency problem of this type of regulator.
Another Alternative In many cases, a high-isolation trans former might be overkill. If you are willing to deal with some noise on your power line and are worried only about large voltage spikes, there is a simple solution.
An MOV (metal oxide varistor) is an inexpensive device designed to clamp voltage spikes. This device acts as an open circuit until a defined voltage is exceeded, at which time it becomes a near short-circuit until the defined voltage drops below the threshold voltage. These devices can absorb huge amounts of energy over a short period of time.
For instance, the General Electric GE-MOV II Series L can absorb up to 50 J of energy, or 6,000A. For power line use, the V130LA series has a varistor voltage of 184V minimum and will safely ignore anything lower than that. The V130LA10A looks like a ceramic disk capacitor (except that it is bright red) and will absorb 30j, or 4,000A.
This is what you find inside most off-the-shelf voltage spike protectors.
When I disassembled the RCA SK400 voltage spike protector, I found a GE V170LA10A MOV that according to the RCA literature, clamps voltages to approximately 435V. The higher rating is probably due to a combined UL/CSA rating, which requires the device to pass 220V AC safely. The MOV is identical to the V130LA10A in capabilities. The SK400 also has a thermal circuit breaker in series with the MOV to prevent the spike protector from causing a fire. Unfortunately, this circuit breaker could open up, and you would have no indication that your protector was inoperative.
Do It Yourself
Adding an MOV to existing equipment is relatively simple, once you understand what you are doing. In many cases, a single MOV across the transformer primary (Fig. 1) is adequate to catch most of the transients. Note that the MOV is after the fuse.
Although MOVs are very reliable, you should protect your equipment from a possible failure that would leave the MOV shorted. Also, if the MOV absorbed a huge spike (such as a lightning strike), the fuse would blow and further protect your equipment.
General Electric has an informative handbook called the Transient Voltage Suppression Manual that goes into great detail about the MOV and provides a lot of useful application data. You can get the booklet from any electronics supplier that carries GE parts.
GE suggests the V130LA10A for low duty cycle consumer applications (e.g., a portable TV), the V130LA20A for medium duty con sumer applications (e.g., a console TV) and the V130LA20B for medium duty light industrial applications (e.g., a photocopier). Using the guidelines in the handbook, you can select which MOV you should use for a specific application.
One positive aspect of using MOVs in power supplies is that you no longer have to worry about transients destroying rectifier diodes. In many cases, you can safely use a rectifier that has a peak reverse voltage (PRV) only slightly higher than the normal peak voltage on the trans former secondary. The MOV on the primary will clip anything that might cause the secondary peak volt age to increase beyond the PRV of the rectifier. While this is not a big deal with small rectifiers (e.g., the IN4000 family), it is important when you are designing a high-current power sup ply for a large power amplifier. Those diodes are expensive.
A single MOV between hot and neutral will protect you from transverse-mode spikes, which are between hot and neutral. Spikes can also develop between ground and the hot and neutral leads. This is called common-mode noise and requires two more MOVs, one between hot and ground and another between neutral and ground. Note that ground must be a good ground connection.
FIGURE 1: Installing an MOV across the primary of the transformer-after the fuse-catches most transients.
FIGURE 2: You can protect your system inexpensively by combining three MOVs and two small chokes on the hot and neutral leads in a box and using multiple outlets.
Otherwise, the energy will not be shunted to “earth.” Adding small chokes on the hot and neutral leads increases power line impedance and tends to lower the peak current absorbed by the MOV. All that is required to provide some impedance is ten turns or so around a ferrite rod from a deceased transistor radio antenna. Wound with #18 AWG or larger wire, it has essentially no loss at 60Hz.
If you were to build a combination of the three MOVs and the chokes in a box and use multiple outlets, you could protect your entire system with a very small investment. Figure 2 shows a complete diagram. Note the fuse in the hot lead protecting your house wiring from a shorted MOV. A fuse is not required in the neutral and ground wires and would, in fact, be unsafe. The reason for this is that if the neutral fuse blew before the hot fuse, you might find yourself completing a path from a hot wire to ground on equipment you would otherwise believe is not energized.
I have also indicated color codes, which conform to NEC standards.
Green is always ground, white is always neutral, and black is the first color choice for a hot lead. (Red and blue are the second and third choices of colors for a hot lead.) While it is not critical to use these specific colors, it is important not to confuse hot and neutral in any wiring. Following a consistent color code helps keep you on the right track. (European readers should replace green with a yellow stripe for ground, blue for neutral and brown for hot.) Be sure the ground connection is a low-resistance ground. If necessary, run an extra wire of at least #12 AWG from the outlet feeding your protection box directly to a cold water pipe or, even better, a 10-foot ground rod driven into the soil. This is important mostly for lightning protection, but might also help save your equipment from damage in freak accidents.
Imagine, for instance, what would happen if a car ran into a power pole in front of your house and the high voltage distribution lines above the house voltage lines on the pole shorted together and suddenly put 4 to 12kV into your house. The MOVs would shunt all this to ground, blowing fuses in the process, but protecting your equipment.
You could spend the rest of your life without ever having a problem with voltage spikes damaging your equipment. But these problems can arise, and for a small investment in parts and time, you can improve the chances of your equipment's surviving typical power line problems.
-- Bill Ruck
Next time, Mr. Ruck will look at noise and other forms of interference from a quality, rather than a reliability, standpoint.
SOURCES:
1. Sola Electric, 1717 Busse Rd., Elk Grove Village, IL 60007.
2. Stancor Products, 131 Godfrey St., Logansport, IN 46947.
3. Topaz Electronics Division, 9192 Topaz Way, San Diego, CA 92123.
4. Transient Voltage Suppression Manual, General Electric Company, W. Genesee St., Auburn, NY 13201.
-------------------
Also see: