Lesson 1 of 29 · 5 min read
Root cause thinking
A fix answers "what failed." Root cause analysis answers "why it failed" and "what lets it fail again."
This course carries "root cause analysis" in its title, so the first lesson says plainly what that discipline is. When a machine goes down, two questions are on the table. "What failed" names the part: the fuse, the seal, the coupling. "Why it failed" names the condition that killed the part. Repair closes the first question. Root cause analysis closes the second, and the second is the one that decides whether you ever see this fault again.
Proximate cause, root cause
A packaging line drops out and you find a blown control fuse. The fuse is the proximate cause: the last link in the chain of causes, the thing that directly took the machine down. Replace it and the machine runs. Look closer and you find the wire that blew it, chafed bare where it rubs a panel edge. That is nearer the root. Look once more and you find the reason it chafes: the strain relief that should anchor that wire was never fitted. That is the root cause. Fuse, wire, missing clamp: one failure, three causes, each one deeper than the last.
Fix at the fuse and the fault returns in weeks. Fix at the wire and it returns in months, because the bare spot re-wears. Fix at the strain relief and it never returns. How deep you fix decides how long the fix lasts.
The failed part is a witness
A part that keeps dying is being killed by a condition outside it. The third fuse in a month did nothing wrong; something keeps overloading it. The third seal in five weeks was fine when it went in; something keeps cooking it. Treat the failed part as a witness: examine it, read the marks on it, and let it tell you what was done to it. The murder weapon is somewhere else on the machine.
Why swap-first costs you
Swapping the failed part sometimes gets the machine running, and it teaches you nothing. You cannot tell whether you removed the cause or reset the countdown, and every recurrence costs the same downtime again, usually on the worst shift. Disciplined diagnosis costs minutes up front and ends the recurrence. Swap only what the evidence points at, and when a part has died more than once, treat the swap as first aid while you hunt the condition that killed it.
Two tools you will use everywhere
The 5 Whys is a walking tool: ask "why" of each answer until the answer is a condition you can remove. Why did the line stop? The fuse blew. Why did the fuse blow? The wire shorted. Why did the wire short? It chafed bare on the panel edge. Why did it chafe? No strain relief anchors it. Why was there no strain relief? It was left off in a rebuild. Five questions took you from a blown fuse to a missing clamp and a rebuild checklist worth fixing. That run of questions and answers is called a Why chain, and the next lesson is about keeping one on the evidence.
A fault tree is a map of candidates: every condition that could produce the symptom, laid out so evidence can eliminate branches one by one. You will use small fault trees constantly on this site. The candidate-cause list in every simulation is one: a set of competing explanations, waiting for your probes to rule them out.
Those are the two tools. The next lesson stays with the 5 Whys long enough to show you where a Why chain goes wrong and how to keep yours on the evidence. After that comes the hands-on loop you will run at every machine: gauge, localize, reason, act. This lesson says what a real fix has to answer. The loop is how you go and find those answers.
Check yourself
Pick an answer to see why it’s right or wrong.