Intermittent failures are difficult because the evidence disappears when the machine recovers. The solution is not to guess faster; it is to capture better data.
Five Principles to Use in the Field
- Define the trigger pattern. Time, temperature, vibration, product type, speed, shift, washdown, and machine position can reveal the condition that creates the fault.
- Use trend and history tools. PLC logs, drive histories, HMI alarms, counters, event timestamps, current traces, and temperature records can capture what humans miss.
- Inspect moving connections. Drag chains, flexing cables, slip rings, plugs, terminals, and vibrating sensors are common intermittent sources.
- Create a safe reproduction plan. If the fault can be reproduced, vary one condition at a time and observe the earliest state change.
- Do not destroy the pattern. Random adjustments can make an intermittent fault less frequent without fixing it, which delays the eventual failure.
Pressure Mistakes to Avoid
- Replacing parts because they are easy to replace.
- Calling the problem fixed after ten good cycles when it normally fails once per shift.
- Clearing logs before extracting them.
- Ignoring environmental clues such as heat, moisture, vibration, or cleaning schedules.
A Practical Breakdown Method
- Record every occurrence with timestamp and conditions.
- Identify common factors across events.
- Choose signals that would reveal the earliest failure.
- Log or trend those signals if possible.
- Reproduce safely or wait with instrumentation in place.
- Verify the repair across a realistic duration and operating range.
Scenario: A machine loses a remote I/O rack several times per week. Power and network checks are normal during daytime investigation. Event times cluster shortly after sanitation. Inspection finds moisture entering a field connector when the area is washed. Improving sealing and routing removes the intermittent network dropout.
Train This Before the Next Callout
Choose one critical machine you know and apply the ideas from this chapter while it is healthy. For intermittent faults: capture before you chase, write down the normal sequence, the key evidence you would want to preserve, and the first three checks you would make if the machine stopped. Then compare your plan with the electrical drawings, PLC diagnostics, maintenance history, and the experience of another technician. A ten-minute rehearsal in daylight can remove several minutes of confusion at night.
3 AM Rule: When the fault will not stay, make the evidence stay.
Chapter Checklist
- Can I describe the symptom without naming an unproven cause?
- Have I preserved the evidence that could disappear after a reset or restart?
- Do I know which system boundary I am testing next and why?
- Have I kept safety controls and temporary changes visible and controlled?
- Can I explain my current facts, hypothesis, and next action in under thirty seconds?