The Fault That Comes Back Bigger: Decision Tree
Why this matters
Some faults do not just repeat, they grow. A small leak becomes a flood, a warm connection becomes a burned bus, a minor noise becomes a seized bearing. The tech who patches the symptom each time it appears is riding a curve that ends in a much larger failure, often a dangerous one, and the customer remembers that you were there twice before it took the whole unit out. Recognizing a progressive fault early, while it is still small, is the difference between a cheap fix and a catastrophe. This tree is how you spot the escalator before it reaches the top.
Safety first: a fault that is growing toward a hazard
Before diagnosis, ask where this fault is heading. If the trend points at a hazard, the escalation itself is the emergency.
- A connection getting hotter each visit is heading toward an electrical fire. De-energize and address it now, do not schedule it.
- A leak getting larger near electrical equipment, a pressure component thinning, a combustion fault worsening: lead with the safety action. Shut it down and make it safe before you study the trend.
A progressive fault that is trending toward a hazard does not get a "watch it" recommendation. It gets stopped.
Start here: is this fault getting worse, or just recurring?
The first cut: distinguish a fault that repeats at the same severity from one that escalates.
- Recurring at constant severity: the same minor fault comes back the same size. Annoying, but stable. Find and fix the cause.
- Escalating: each occurrence is worse than the last, or the intervals between them are shrinking. This is the dangerous pattern. Something is degrading, and the degradation feeds itself.
Ask the customer or read the history: is it bigger than last time, or more often? Either answer means escalating.
Why faults escalate: the feedback loop
Progressive faults grow because the fault damages the thing that would have contained it. The damage makes the next fault worse, which does more damage. Common loops:
- Heat: a high-resistance connection warms, the heat oxidizes the contact, the worse contact raises resistance, which makes more heat. It runs away toward a burn.
- Leak and corrosion: a small leak corrodes the surrounding metal, the corrosion enlarges the leak path, which wets more metal.
- Vibration and wear: a small imbalance accelerates bearing wear, the worn bearing increases the imbalance, which wears it faster.
- Pressure and fatigue: a stressed component flexes, the flexing fatigues the material, the weaker material flexes more.
When you can name the loop, you know it will not hold still. A self-reinforcing fault has no stable plateau; waiting only raises the stakes.
Read the trend, not just the snapshot
A single reading tells you the state now. The escalation is in the change over time.
- Compare to last visit: is the measured value (temperature, leak rate, vibration, current) worse than the documented prior reading? This is exactly why recording readings matters, see the related documenting article.
- Compare the intervals: failures every few months, then every few weeks, then every few days, is an accelerating curve even if each event looks small.
- Look at the collateral: a fault that is enlarging the damage around it (a growing burn mark, spreading corrosion, more debris) is escalating regardless of what the live reading says today.
If you have no prior data, take careful readings now and tell the customer the next visit will confirm the trend.
Decide: fix the cause, or it will be back bigger
| Situation | Action |
|---|---|
| Escalating toward a hazard | Stop it now, safety action first |
| Self-reinforcing loop identified | Fix the cause this visit, do not patch |
| Worse than documented prior reading | Treat as progressive, find root cause |
| Constant severity, no growth | Fix cause at normal priority |
| Cannot confirm trend, no prior data | Record readings, schedule a confirming visit |
The trap is patching the symptom of a progressive fault because each individual occurrence looks minor. Each one looks minor right up until the one that does not. When you have identified a feedback loop, the only real fix is breaking the loop at its cause, not replacing what it most recently damaged.
Tell the customer the trajectory
A progressive fault is the one where you must be explicit about where it is going. "This is small now, but it is getting worse each time, and the way it fails is the part of the system gets damaged, which makes it fail faster. If we only patch what broke this time, it comes back bigger. Here is the fix that stops it." Document the trend and the readings so the trajectory is on record.
The recap
- If the fault trends toward a hazard, stop it now, safety action first.
- Separate recurring-at-constant-severity from escalating.
- Name the feedback loop. A self-reinforcing fault will not hold still.
- Read the trend: compare to prior readings and shrinking intervals.
- Fix the cause; patching the latest damage just resets the clock.
- Tell the customer the trajectory and document the readings.
References
- Trade-standard practice on progressive and self-reinforcing failure modes
- OSHA and NFPA guidance on escalating electrical and pressure hazards
- Manufacturer documentation on degradation and failure progression
- See related: Reading Rust and Corrosion Patterns; Documenting an Unsolved Fault for the Next Tech (how-to); The Reset That Masks the Problem (decision tree)