Simple power quality problems often evolve into tangled webs that take a long time to unravel. As commercial and industrial systems become more integrated and rely more heavily on bidirectional information from multiple sources that are powered from different supplies, it can take considerable time to understand the interdependencies of how one part acts and reacts to the others. Which piece of equipment experiences a problem first? Did that equipment still continue to operate properly after the event but sent incorrect information to other equipment that took the wrong action? In which path did the evil come? Was it the power system, backup power system, communication circuits, sensors or all of the above?
Redundant power systems don’t mean perfect uninterruptible power. Case in point is one large manufacturing facility that went from several large uninterruptible power supply (UPS) systems feeding critical parts of each production line to thousands of smaller UPS units powering sections of each production line. The thinking was the classic, “don’t put all the eggs in one basket,” but they forgot the corollary. With that many eggs, the odds of having a broken one are much greater. Instead of a problem once or twice a year, failures occurred on a weekly basis, shutting down a line.
As another recent example, in a residence, the lights on a chandelier occasionally blinked, a common complaint. It didn’t cause any real problem, no financial loss and no epileptic seizure from the flicker. It was just annoying when eating dinner. The homeowner wanted it to go away. The first questions asked in a power quality investigation are usually what was affected, when was it first noticed, what changed from before the problem showed up to now, and is anything else affected. The homeowner noticed it after remodeling the kitchen, but the flicker may have started after the solar panels were put on the roof. No clocks went blinky when the lights dimmed. There may have been a sound coming from the basement at the same time.
Typically, such dips in lighting are the result of a large load starting up, causing a voltage drop. Given the sound association, it could have been a heat pump (but they didn’t have one); maybe the refrigerator, coffee pot, or microwave cycling (they claim none were on); or a sump pump (something in the basement that makes noise). A connection to the solar panels didn’t seem likely, as they aren’t doing much at night when the lights are on in the dining room. Chandelier bulbs, with their small filaments, do not have much photon inertia, so a short reduction in voltage would be quite noticeable in their light output. Dimmer switches also make flicker more noticeable, as even less thermal/photon energy is provided to the bulbs.
Determining what else is on the same circuit was the next step. After finding the right breaker and switching it off to see what else was on the circuit, we found none of the aforementioned possible offenders were on that circuit. While it is possible for a sag to be caused by a large load on another circuit that brings down the voltage on the entire bus, it is much more common for one on the same circuit to be the source of the problem. Instead, what was found was more intriguing and a bit scary.
The picture above shows the wires in the receptacle box where the dimmer switch for the light was located. Two metallic two-conductor cables came in the bottom of the box, and the conductors were appropriately twisted together. A metallic cable from the top had only the white conductor connected to another white conductor; though, in the picture, the black and white are shown individually capped. Into the back of the box (without a wire clamp), were two nonmetallic, three-conductor cables. One cable was traced to go to another receptacle box that controlled recessed lights in the adjacent living room. The other one had the red conductor going to one leg of the dimmer switch for the light in question. The other wire of the dimmer switch went to the junction of the aforementioned hot conductors. The black conductor (thinking hot) of the three-conductor cable was capped off, while the white (thinking neutral) conductor was originally in the same wire nut (but shown exposed in the picture) as the white wire coming out the top. The white conductor had 127 volts on it! There was no black tape or marking to indicate it was being used a return leg of a switch.
It turned out this white wire was powered from another breaker marked as “subpanel,” which fed a subpanel for the second floor bedrooms, so the light in question was fed from a cable with different breakers related to the wires within the cable and had the hot and return from different circuits. Obviously, a “qualified person” didn’t create this situation and is another example of why homeowners shouldn’t dabble in what a licensed electrician is trained to do properly. Proper wiring resulted in another not-so-simple power quality problem solved.