An equipment trip that happens once a week and then disappears for days is one of the hardest problems an electrical team can face. Nothing looks wrong during a routine walk-through, yet something clearly triggers the fault when nobody’s watching.
Closing that gap takes a structured approach to power quality troubleshooting, one that captures the right data at the right moment rather than guessing at a cause after the fact. This article lays out a workflow for chasing down that kind of elusive fault, from the first symptom report through to a defensible conclusion.
Start With the Symptom and Operating Context
Before any meter gets connected, the story behind the fault deserves careful gathering. The three checks below turn a vague complaint into something an investigation can actually work with.
Document When Resets, Trips, or Malfunctions Occur
Ask exactly when the problem shows up: time of day, day of the week, and any pattern in how the symptom presents itself. A drive that fails every afternoon points somewhere very different than one that trips at random, so this timeline becomes the backbone of everything that follows.
Identify Loads, Switching Events, and Recent System Changes
Large motors starting, capacitor banks switching, or nearby equipment cycling on and off can all disturb a shared supply. Recent changes matter too, since a new piece of equipment, a modified circuit, or a utility-side alteration often lines up suspiciously well with when the intermittent electrical problems first began.
Determine How Often the Symptom Occurs and How Long to Monitor
Frequency shapes the monitoring plan directly. A fault occurring daily might only need a day or two of logging, while one occurring once a month calls for extended power quality monitoring, sometimes running for several weeks before enough events accumulate to draw a reliable conclusion. Where frequency itself is unclear, starting with a shorter trial run and extending it based on what turns up is often a practical choice.
Plan Where and How to Measure
Rough as it may be at this stage, the pattern already points toward where a meter should go and what kind of instrument the job calls for. Turning that rough picture into an actual measurement plan means settling a few practical questions before anyone touches a panel.
Choose Measurement Points Based on the Affected Equipment and Distribution Path
Measuring as close to the affected equipment as safely possible usually gives the clearest picture, though a point further upstream can help separate a local wiring issue from a supply-side disturbance. Mapping the distribution path between the suspected source and the symptom location guides this choice before any connection is made.
Confirm the Instrument, Sensors, Measurement Range, and CAT Rating Suit the Measurement Environment
A power quality analyzer needs to match the environment it’s dropped into. That check typically covers:
- Voltage and current ranges wide enough to capture the expected supply, with headroom for unexpected excursions.
- Current sensors or clamps sized correctly for the conductors involved.
- A CAT rating appropriate to the panel or bus being connected to.
- Enough onboard memory or logging capacity to cover the planned monitoring period.
Confirm Qualified Personnel and Safe Connection Procedures
Connecting monitoring equipment to a live panel carries real risk, so only qualified personnel following proper lockout, PPE, and safe work procedures should make the connections. Skipping this step for the sake of speed defeats the purpose of an investigation meant to prevent problems.
Capture the Electrical Conditions Around the Event
With the instrument in place, the focus shifts to actually collecting usable data across the monitoring window. Getting a picture worth trusting means paying attention to several distinct layers of that data beyond the headline numbers alone.
Record Voltage and Current Trends Over a Representative Period
Baseline trends matter as much as the fault itself. A representative logging period, covering normal operation alongside whatever triggers the symptom, gives the investigator something to compare an anomalous reading against rather than judging it in isolation.
Capture Sags, Swells, Interruptions, Transients, and Inrush Events Where Relevant
Depending on the suspected cause, several disturbance types deserve specific attention:
- Voltage sags – brief drops often tied to motor starting or fault clearing elsewhere on the system.
- Voltage swells – momentary rises frequently linked to load rejection or switching events.
- Interruptions – complete loss of supply, however short.
- Electrical transients – fast spikes typically caused by lightning, switching, or capacitor energizing.
- Inrush current – the surge drawn when a motor or transformer first energizes, which can masquerade as a fault if unaccounted for.
Check Harmonics, Unbalance, Frequency, and Power Parameters When the Symptom Suggests Them
Where the symptom points toward heating, nuisance tripping on protective devices, or drive misbehavior, broader parameters warrant a look. Harmonic distortion often surfaces as overheated neutral conductors or transformers running warmer than expected, commonly traced back to variable frequency drives or other nonlinear loads on the same feeder. Voltage unbalance shows up differently, typically as a motor running hotter on one phase or drawing uneven current across the three legs.
Frequency deviation matters mainly on sites running local generation, where a drifting frequency can trip sensitive equipment before voltage readings look unusual at all. Power factor readings round out the picture, and reviewing all parameters together, rather than checking each in isolation, makes it easier to spot which combination actually lines up with the reported symptom.
Keep Measurement Time References Aligned With Equipment Alarms or Incident Logs
An analyzer’s internal clock should be synchronized with the equipment’s own alarm log or building management timestamps before monitoring begins. Without that alignment, correlating a recorded disturbance with an actual reported trip becomes a matter of guesswork rather than confirmation.
Correlate the Recorded Data With Equipment Behavior
A log full of recorded events is only raw material until it gets connected back to what actually happened on the equipment side. Making that connection reliably comes down to a few habits of comparison.
Use Voltage and Current Changes to Assess Upstream and Downstream Clues
A disturbance visible upstream of the affected equipment suggests a shared-supply cause affecting other loads too, while one appearing downstream points toward something local to that specific circuit or device. Comparing readings taken at multiple points helps narrow this distinction considerably.
Review Event Magnitude, Duration, Timing, and Repetition Instead of Isolated Peaks
A single sharp reading rarely tells the whole story on its own. A far more reliable picture comes from weighing several dimensions of an event together:
- Magnitude – how deep the disturbance goes relative to nominal levels.
- Duration – how long the event lasts, since a fleeting dip and a sustained one carry different implications.
- Timing – when the event occurs relative to known switching or load activity.
- Repetition – whether a similar pattern shows up across multiple occurrences rather than just once.
Distinguish an Electrical Anomaly From a Confirmed Root Cause
A recorded disturbance confirms that something unusual happened electrically at a given moment. It does not automatically confirm why, and treating correlation as proof without further verification is a common way investigations go off track. The recorded event narrows the search, and it doesn’t necessarily end it.
Decide What to Verify Next
Correlated data narrows the list of suspects, but narrowing isn’t the same as confirming. Closing that remaining gap calls for a short round of checks before anything gets written into a final report.
Recheck Measurement Setup and Data Quality Before Drawing Conclusions
Before treating any result as final, revisit the setup itself. Loose current clamps, an unsynchronized clock, or a sensor sized incorrectly for the conductor can all produce data that looks meaningful but actually reflects a measurement flaw rather than a genuine electrical event.
Compare Findings With Equipment Requirements, Project Criteria, and Manufacturer Guidance
Once the data holds up under review, compare it against relevant benchmarks:
- The affected equipment’s published voltage tolerance and ride-through capability.
- Applicable project specifications or contractual power quality limits.
- Manufacturer guidance on what disturbance levels the device is rated to withstand.
Know When Additional Electrical Tests or Longer Monitoring Are Needed
A single monitoring window sometimes isn’t enough to catch a genuinely rare event. In other cases, the data collected points toward a mechanical or thermal cause that electrical power quality logging alone can’t confirm.
Extending the monitoring period is one reasonable next move. Adding insulation or thermal testing, or involving the equipment manufacturer directly, are other options worth considering when the picture remains incomplete.
Reading the Pattern: Turning Scattered Trips Into a Solved Problem
An intermittent fault rarely announces its cause in a single measurement. Chasing one feels less like finding a smoking gun and more like assembling a pattern from scattered fragments over time. The workflow above works precisely because it treats each stage as a filter, narrowing possibilities rather than jumping to a verdict from one convincing-looking spike. That discipline, more than any single instrument reading, is what turns a maddening intermittent problem into one that stays solved.