Sometimes, a power quality investigation can lead you to see things that you aren’t really looking for and might be a distraction from what the customer wants you to do. But there is a point where you may also feel an obligation to tell the customer that there are more potential problems in their electrical system than just the obvious ones that you were called in to fix. A recent investigation at a residential dwelling falls into that scenario.
Owners of a two-story colonial-style house built in 1999 had recently added a 24-by-14-foot sunroom off the kitchen. The electrical work included eight receptacles, four exterior luminaires, and eight recessed luminaries, along with a 60-ampere (A) line for a new heating, ventilating and air conditioning (HVAC) unit for just that room, with electric heat.
During the testing of the installation, the electrician measured 135 volts (V) on both phases of the split-phase service. The next morning, a secondary test with a different meter found both phases to be around 124V, but, as a precaution, a power quality monitor was installed to ensure it didn’t recur. The high reading turned out to be a low-battery condition with the electrician’s digital volt meter, but there was some interesting data that showed an 8V voltage in one phase every 34 minutes, day and night, under different outside temperature conditions. It happened during a week when the owners were on vacation. Since the original threat of a dangerously high steady-state voltage level proved to be a nonissue, the owners weren’t interested in finding the source of the potentially problematic periodic voltage drops.
Fast-forward two months to when the room was being used over the Thanksgiving weekend. The new arc-fault circuit interrupter (AFCI) breaker for the lighting circuit tripped twice while most of the house was fully lit, inside and out. In one instance, it tripped when a young child was rapidly flipping the switch that controlled six of the 65-watt (W) halogen recessed lights, so maybe it tripped on arcs from the rapid interruption or maybe from some other cause. The power quality monitor was brought back and re-installed; this time it monitored the voltage and current on both phases in addition to the current on the circuit that tripped.
According to the homeowner, nothing happened for the two-week period that the monitor was installed, since the breaker didn’t trip. The data was retrieved and analyzed. Figure 1 shows a timeplot of both phase voltages and the neutral-ground (N-G) voltage. Figure 2 shows both phase currents and just the circuit that had previously tripped.
Since the monitor was at the breaker panel, there shouldn’t be any N-G voltage if the bond is made correctly, but it is worth measuring when you have an eight-channel monitor. The 200A service wasn’t overloaded, and there was only one recorded sag, but it was a minimal, one-cycle sag and didn’t affect any loads. Though the circuit didn’t trip while being monitored, the time plots show the 15A level was exceeded often during the monitoring period. The duration above 15A wasn’t very long—less than 10 seconds—but it still was a problem. The current’s waveforms when this occurred turned up some unexpected results for what was supposed to be a lighting-only circuit. The on-off pattern of the current was indicative of a microwave. Further investigation found that most of the peaks occurred during lunch and dinner time and that the kitchen was improperly wired; the microwave and dishwasher were on the same kitchen lighting circuit; the new room’s lighting circuit also was connected to it.
Some other interesting data, such as the current waveforms and rms plot during startup of the HVAC system, showed up. Figure 3 wasn’t seen until the seemingly “normal” voltage rms plots were examined more closely, revealing a similar pattern to the prior 34-second, 8V voltage-drop pattern seen three months earlier when it was warm enough for the AC to be on. Though now lasting 30 seconds and with only a 2–3V drop, the drops still occured all day and all night, at the same depth and interval, regardless of the winter weather conditions or occupancy use. With the current being monitored this time, it is easy to see that the cause of the problem is load side, since the current increases significantly (20A) with each drop. Using Ohm’s Law and the formula for approximating source impedance (V2 – V1)/(I1 – I2), it shows an acceptable source impedance of 0.1 ohm.
Once again, the homeowners’ interest stopped at determining that the microwave needed its own circuit to prevent the “lights out” during the dinner parties.
The breaker did trip two more times during Christmas dinner, after the monitor was removed and before the electrician had a chance to come back.