Skip to main content

Lesson 10.3: Root Cause Analysis of Mechanical Failures


Technical Context

Replacing a broken part may get the robot running, but it does not explain the failure. If the load, geometry, or assembly error remains, the replacement faces the same risk.

Root cause analysis is the discipline of finding out why the part was carrying a load it could not survive.


The Broken Part Is Evidence

Before replacing anything, look at the failure itself. Different causes leave different marks.

What you seeLikely cause
Clean break with no deformationBrittle failure, often impact or a stress concentration
Bent and stretched before breakingOverload, the part was undersized for the load
Break at a sharp internal cornerStress concentration, needed a fillet
Break along printed layer linesWrong print orientation for the load
Threads stripped, screw intactInsufficient thread engagement, or a stripped soft material
Screw shearedThe joint was carrying load in shear rather than by clamping
Elongated holeThe joint was loose and the part shifted under load
Polished or worn surfaceRubbing that was not intended, a bind or misalignment
Discoloration or melted plastic near a motorSustained overcurrent, likely a stall condition

Photograph the broken part before doing anything else. It goes in the notebook, and the photograph is often the only evidence left after the repair.


Ask Why Repeatedly

Ask why repeatedly until the chain reaches a cause the team can change.

The slide stopped moving.
Why? The motor stalled.
Why? The slide was binding at full extension.
Why? The second stage rail was not parallel to the first.
Why? The mounting plate flexed when the slide was loaded.
Why? The plate is 1/8 in polycarbonate and spans 9 in
unsupported.

The first answer, "the motor stalled," describes the symptom. The final answer identifies the flexible plate that allowed the slide to bind and points to a relevant design change.

Stop when you reach something you control

The chain ends when you reach a design decision you can change: a material, a dimension, a support, a fastener count. If the last answer is "because the robot got hit," keep going, since being hit is a normal condition and the design should survive it.


Distinguish Three Kinds of Failure

Design failure. The part was never adequate for the load it sees. Recalculation is needed, not a stronger version of the same part.

Build failure. The design was adequate and the part was made or assembled incorrectly: a fastener not tightened, a spacer omitted, a misaligned bearing. The fix is a process change, such as a checklist or a stack-up recorded in the notebook.

Wear failure. The part was adequate but has a finite life: compliant wheels, surgical tubing, chain, bearings. The fix is a replacement schedule and spares in the pit, not a redesign.

Classify the failure before choosing a response. Redesigning a normal wear item may add unnecessary work, while treating a design failure as routine wear leaves the cause in place.


The Fix Must Address the Cause

Once the root cause is identified, check that the proposed fix actually removes it.

Root causeFix that addresses itFix that does not
Plate flexes, causing bindStiffen or support the plateBigger motor
Insufficient thread engagementNut, insert, or thicker materialMore torque on the screw
Sharp corner concentrates stressAdd a filletThicker part with the same corner
Joint loose, hole elongatingSecond fastener, tighten properlyLarger screw in the same loose joint
Motor overheats holding a loadCounterbalance or brakeReplace the motor

The fixes in the right column address the symptom without removing the listed cause.


Record It

A failure analysis entry should contain the symptom, evidence including a photograph, the why chain, the identified root cause, the fix, and verification that the fix worked.

That last part matters: after the fix, re-run the test from Lesson 10.2 and record the new number. A fix that was never verified is a hypothesis.


Fill-in-the-Blank Practice

  1. A break at a sharp internal corner with no deformation suggests a stress __________.
  2. The why chain should continue until the answer is a design decision the team can __________.
  3. A failure caused by a part that has a finite life, such as surgical tubing, is a __________ failure and calls for a replacement schedule rather than a redesign.
Show answers
  1. concentration
  2. change (control)
  3. wear

Exercise

Take the most recent thing that broke on your robot. Photograph it if you still have it, then write the why chain until you reach a design decision. Classify it as a design, build, or wear failure, and check that the fix your team applied actually addressed the root cause.

Ready to move on?

Only mark complete if you genuinely understand the material. Your progress will be saved in this browser.

Stuck on this lesson?

Ask about anything on this page. It can see which lesson you have open and which part you are reading.