How to Set Up a Monitor to Catch a Fault You Can't Stay For

Why this matters

Some faults hit at three in the morning, or once a week, or only during a storm, and no one is paying you to sit and wait for them. A monitor left in place watches on your behalf and reports the event with a timestamp and a value. That turns the next visit from another fruitless look into a targeted fix. The skill is choosing the right thing to watch, the right tool to watch it with, and placing it so it captures the fault itself and not just the aftermath.

Step 1: Pick the one parameter that moves when the fault hits

You cannot watch everything, so watch the thing that changes at the moment of failure. If the fault is a voltage sag, monitor voltage. If it is an overnight freeze-up, monitor temperature. If it is a pressure spike, monitor pressure. Choosing the parameter that actually moves with the fault is most of the setup.

Step 2: Match the tool to the event

Reach for the recorder that fits:

• A min/max meter that holds the highest and lowest it saw while you were gone.
• A data logger that records the parameter at intervals with timestamps.
• A recording gauge for pressure or temperature over time.
• A camera pointed at a display or an indicator light for a visible event.
• The system's own onboard log or alert feature, if it has one.

Step 3: Place it to catch the fault, and do it safely

Set the monitor where and how it will see the actual event, with the trigger or threshold set to capture it rather than miss it just under the line. Safety first: leaving equipment connected to a live or energized system means securing it so it cannot fall, short, or become a hazard while unattended, using leads and clamps rated for the circuit, and labeling it so no one mistakes it for a fault or disconnects it blind. Never leave a monitor rigged in a way that defeats a guard or a safety device.

Step 4: Set duration and sampling to match the fault

A logger sampling too slowly will step right over a brief event; one sampling too fast fills its memory before the fault ever happens. Set the interval fine enough to catch the event and the run length long enough to reach the next occurrence. Base both on what you already know about how often and how fast the fault appears.

Step 5: Record the baseline before you leave

Note the normal readings while everything is behaving. Without a baseline, a captured number means nothing, you cannot tell a spike from the ordinary. With one, any deviation reads clearly as the event. A quick set of normal values before you walk out is what makes the captured data interpretable.

Step 6: Retrieve, read, and place it in time

Come back, pull the data, and line the captured event up against everything else that was happening. Feed the timestamp into a timeline and cross-check the conditions. The captured event plus its context is the diagnosis the monitor was there to get, and a lone number with no context is only half of it.

References

• Trade-standard practice for unattended monitoring and data logging
• OSHA and NFPA 70E guidance on securing test equipment on energized systems
• See related: How to Log an Intermittent Fault Over Days; How to Build a Timeline of When a Fault Appears