You finally were able to buy a power quality monitor to determine what is causing your customers’ equipment to misoperate. From the frustration of waiting for Windows to reboot to the inconvenience of cleaning out solidified plastic pellets when the extruder stopped extruding, power quality phenomena problems range from the annoying to the expensive. But as an electrical contractor, the return on investment of the power quality monitor purchases isn’t in the increased productivity of the customer’s facility but rather on your ability to find the answers quickly and move on to the next paying job.


In the past, I have covered rules of thumb on what the collected power quality data means, what the possible sources of the problem are, and how to mitigate the problem, including finding the right location to monitor, the right things to monitor, and the right monitor for the job. Hopefully, my March article helped you answer those questions. But how long to monitor, how to set it up, and how to know if you have captured what you need are equally important questions.


The length of the monitoring period depends on the objective. If it is to record a baseline survey, data should be collected for one business cycle (how long it takes for the process to repeat itself at the facility). A factory that has the same operation 24/7 would have a one-day cycle. A single-shift facility that doesn’t work on weekends will usually have different loading on Monday morning and in the evening/night time, so a week is a typical monitoring period. If the purpose is a troubleshooting monitoring program, the time it takes to reproduce the problem at least once, preferably twice, is the minimum. My shortest monitoring period was 2 minutes, which gathered enough data to show that the bearing failures were caused by very high harmonic levels from old direct current drives.


How best to set up the monitor is somewhat dependent on the capabilities of the monitor itself, as Figure 1 shows. Automatic setups, which some monitors have, may only be useful to get a quick half-hour baseline to better determine how to more precisely set the thresholds. Setting the event thresholds too close to typical voltage and current excursions will give you so much data that it can result in analysis paralysis. More typical is to set the thresholds a little narrower than the susceptibility limits of the equipment on the circuit. For example, if the power line communication system trips offline at 70 percent of nominal, setting the thresholds to 80 percent of nominal will let you gather some events that shouldn’t impact the equipment and some events that likely will do so. 


If in doubt, start with the IEEE 1159 typical settings of 90 percent, 110 percent and 10 percent of nominal for sags, swells and interruptions, respectively, as Figure 2 shows. Incorrectly setting up current triggers commonly results in memories filled with useless information. While the voltage is typically stable within narrow limits except during power quality disturbances, the current levels and other related parameters, such as watts, power factor and volt-amperes-reactive, often have swings of 5-to-1 during a business cycle.


And it is not just the depth of the sag that causes problems. The duration is often equally as important, depending on the type of power supply in the equipment. Think of reduction-in-voltage multiplied by time as missing energy. A deep sag for a short duration may have the same result as a shallow sag for an extended duration. That means setting the duration of the pre/post cycles of voltage and current should be relative to the load’s susceptibilities. Again, if in doubt, setting eight pre and 32 post cycles is a good starting point.


Most instruments can periodically save or journal hundreds if not thousands of parameters at rates down to once per second. Don’t just set things to the maximum. Why take home the entire ocean when all you want is the fish? Processing gigabytes of recorded data is perhaps the hardest task, and it can be extremely slow if it has to be retrieved remotely using modems. Having a top-notch power quality monitoring software program can make life much easier. Be wary of automatic report generators that crank out data into templates, resulting in a lot of paper without accurately characterizing the system and rarely finding the problem. Also, remember that the reader of the report may have much less power quality knowledge than you, so putting in a lot of “interesting” waveforms without detailed explanations isn’t likely to help anyone.


Getting familiar with these and the aforementioned rules will have you off to the races in no time, and you can get that return on investment on your monitor purchase in less than a year. It’s just more money in your pocket and even more to your happy and productive customer’s bottom line.