A great deal of time is spent defining, identifying, and discussing power quality relationships. These activities fall into the "book learning" aspect of site analysis and are necessary in understanding power quality problems. However, nothing beats seeing the problems firsthand; then, you can correlate the book learning with the practical. In other words, you then can tie together the principles with the required practices.
Practical site analysis guidelines
There are some practical guidelines involved in conducting a typical site analysis. They are based on frequently found disturbances and their characteristics.
Mystery disturbances. First, examine a site for what we call "mystery" disturbances: those upsets to a process or sensitive equipment operation that don't seem to correspond to any identifiable source of power disturbance. Such things as ground loops, high speed transients, lightning, and common mode electrical noise come to mind. Many of these events are here and gone in such a short time frame that they are not easily identified, except with a power disturbance analyzer using highspeed wave shape or event capture.
Repetitive, cyclical disturbances. Secondly, you should look for those disturbances that do have a connection with the power distribution system, both within and outside of a facility. These problems will be repetitive and cyclical in nature, definitely power-related, and line-to-line. Examples include voltage sags and surges, momentary interruptions by circuit breaker operations, and power interruptions.
We must know which disturbance we are facing; looking for a transitory disturbance with tools specifically made to indicate repetitive ones may not give us the right answers.
Harmonic distortion. Thirdly, we should look for those disturbances related to the integer multiples of the fundamental power frequency (60 Hz), the area we call harmonic distortion. We know that this area is a subset of the power related area, since these harmonic currents and voltages are recurring. However, we may need special tactics in searching out our problem and identifying our solution alternatives.
Remember the basics about 3-phase harmonic producers (phase-to-phase nonlinear loads such as VFDs and UPSs) and single-phase switching devices (phase-to-neutral nonlinear loads such as PCs and electronic ballasts). The current spectrum we see and its effect on system voltages will be indicators of the dominant harmonic present.
Typical mystery disturbance scenario
A site is bothered by frequent interruptions that seem to occur when nothing is happening on the facility's electrical distribution circuits. Typically, the equipment goes down on a "blue sky" day, when there is not a utility problem for hundreds of miles. It's easier to understand the disturbance if it occurred during an electrical storm, or when high winds might be blowing the power lines together. But in this case, nothing is happening.
Typically, the story goes something like this. When an interruption occurs one morning, the operations manager calls the customer service engineer for the equipment in question, who then comes to the facility to analyze the problem. This engineer checks the hardware, software, and operating system and finds nothing amiss. He or she may even comment to the owner or manager that there must be an "environmental" problem, since all of the equipment systems are OK.
In the past, we would be stupefied by this answer. Now we're more alert and immediately suspect a ground noise problem. In fact, the history of mystery disturbance problems over the past 15 to 20 years shows that improper facility wiring and grounding are at the heart of most of these problems.
Suppose the owner has heard the same "environmental" answer a number of times and thus wants a more definitive answer. To come up with a "better" explanation for the trouble, the customer service engineer cites a "defective" circuit board, removes it from the affected piece of equipment and notes in the room log that this element is replaced. The owner now has a more confident feeling that some one knows what caused the disturbance.
What happens when the same type of disturbance occurs again? Don't be surprised if the customer service engineer removes the same circuit board and replaces it with the very one removed the time before. And on and on it goes.
Mystery disturbance case history
We performed a study on a small section of an office where sensitive process equipment was installed; this equipment was operating erratically. At this site, the process equipment vendor had called for a motor-generator (MG) set to "buffer" and isolate the power coming to the process. Nevertheless, the process was still being interrupted. The customer service engineer was having a difficult time trying to explain to the owner what was happening.
When we examined the wiring, we found that it was installed properly. However, the electrical grounding system was another matter; it allowed noise paths found in the building to combine with the sensitive process signal. This electrical noise would increase to a level several volts higher than the process signal voltage. When this happened, the process would shut down.
The owner thought the MG set was causing the problem and wanted it removed. This reasoning was based on the viewing of phase voltage wave shapes of the power circuit. The mistake was that power interactions, not signal disturbances, were being shown.
We were able to demonstrate to everyone's satisfaction that the MG set was, in fact, doing its job and that the problem was coming from improper ground connections. This was done by connecting a disturbance analyzer between the power system neutral and the case ground of the process. Both the owner and the customer service engineer assumed that this point had the same ground potential, since that's the way it was shown on the drawings.
Remember, not all drawings reflect what's actually installed. Many times we're misled by the ground symbol itself. The abundance of this symbol on drawings leads us to believe that all such connections have the same contact with the earth reference potential, that they're at equipotential with the power ground.
If we can be sure that there are no ground loops, or at least correct and/or improve any suspect wiring, then we can be certain that ground noise is not the culprit.
Mystery disturbance analysis procedure
First determine if, and/or where, the poorly operating system has a ground loop (a connection where the signal ground and the power ground are attached to ground potential at two different points). This is done by taking voltage measurements with a power disturbance analyzer between the power system neutral at a panelboard and the case ground of the equipment in question. Remember, we're not looking for the ground in the panel; we're looking for the point of signal reference for the device we suspect has a common mode problem. This leads us to the second step.
We then should look at those devices specified with "clean," "private," "dedicated," "special," or other specially named ground. (For many years, telephone switches, computers, central processing units (CPUs), and computer numerical control (CNC) machines all had these special "separate" ground connections, all in violation of the NEC. And, all these points were subject to ground looping.) Corruption of the process signal occurs when the impedance between the two points results in a voltage differential, which drives a "loop" of potential right through the driver and/or receiver circuits in the signal path.
With the analyzer connected between the neutral and case ground, we'll be able to see if there are multiple connection points, or if the points have the same potential. If the wires from our two points go to the same connection point, our analyzer will show little "surge" between them. If, however, the points are at different potentials, a steep wave front signature of a "spike" will be seen.
You also can test for this condition with a simple impedance meter. The diagram shows how and why this works. First, shut down the problem system and, disconnect the respective device dedicated ground connections at their source(s) and connect them together in open air. The impedance of this ground connection setup should be infinite because we have an open air gap between the open air connection point and our supposed single-point ground. If our meter reads zero impedance, then a hidden connection exists around the back of the equipment, providing another pathway to ground. What we thought was a ground reference radial, or direct connection, is obviously not the case in reality but rather a daisy-chained ground.