The Repair LibraryRead · Learn · Master

Failure Analysis — Purpose and Process

Almost everything this handbook has taught up to now has aimed at one thing: find the broken part and replace it. This chapter opens a discipline that goes a layer deeper and asks the question replacement alone never answers — not what failed, but why — because a repair that swaps a failed component without discovering the reason it failed has not fixed the fault, it has only reset the clock on it. Failure analysis is the systematic investigation of why a component or a system failed, and its purpose for a repair technician is intensely practical: a blown part is very often the victim of something else, and a new part dropped into the same conditions meets the same fate, so the difference between a repair that lasts and one that fails again next week is whether the cause behind the symptom was found. The section's central idea is that this investigation is a process, not a guess — a repeatable, evidence-based method that anyone can follow to a defensible conclusion rather than an intuition that works only for the technician who happens to have it. That process moves along a chain the section makes explicit. It begins with the failure mode, the way the failure shows itself — an open, a short, a burned package, a cracked joint, a drift out of tolerance. It works inward to the failure mechanism, the physical process that actually produced that mode — the thermal runaway, the dielectric breakdown, the electromigration, the solder-joint fatigue that is what physically happened to the part. And it traces back to the root cause, the originating reason the mechanism was set in motion — the overstress event, the thin design margin, the manufacturing defect, the wear of age, the environment — which is the thing a repair must actually address. The section frames the disciplined steps that walk this chain — preserving the evidence before it is disturbed, characterizing the failure, determining the mechanism, tracing to the cause, and closing with a corrective action that fixes the cause rather than the symptom — and it sets up the rest of the chapter, which takes root cause analysis, the destructive and non-destructive toolkit, and the written report each to its own depth. The point it leaves the technician with is the one the whole chapter is built on: a failure is a question, and analysis is the discipline of answering why, because only the answer makes a repair permanent.

AdvancedLow Risk23 min read

What You Will Learn

  • You will learn what failure analysis is and why finding why a part failed, not just that it failed, makes a repair permanent.
  • You will learn the difference between the failure mode, the failure mechanism, and the root cause, and how analysis traces the chain between them.
  • You will learn that failure analysis is a repeatable, evidence-based process rather than a guess.
  • You will learn the steps of that process — preserve, characterize, determine the mechanism, trace to the cause, and correct.
  • You will learn why a corrective action must address the root cause, not the symptom, or the failure returns.

What You Will Be Able To Do

  • You will be able to explain the purpose of failure analysis and why replacement without a cause is an incomplete repair.
  • You will be able to distinguish a failure mode from a failure mechanism from a root cause.
  • You will be able to follow the failure-analysis process as repeatable, evidence-based steps.
  • You will be able to preserve the evidence of a failure before it is disturbed.
  • You will be able to frame a corrective action that addresses the root cause rather than the symptom.

Required Tools

  • The as-found record and camera — failure analysis begins by preserving the evidence before anything is disturbed
  • A microscope and good light — the failure mode and its physical signature are read at magnification
  • A notebook structured as a failure chain — mode, mechanism, root cause, and corrective action recorded as the analysis moves along it
  • The board's context — the circuit, its history, and its conditions, because a failure's cause is often outside the failed part

When NOT to Attempt This

Do not attempt this section if any of the following apply to you:

  • You are not comfortable working with small surface-mount components.
  • You have not completed the prerequisite sections for this skill.
  • You do not have the required tools in working condition.

Section Overview

Most of this handbook has taught finding the broken part and replacing it; this chapter asks the deeper question — not what failed, but why (the-troubleshooting-process). A repair that swaps a failed part without finding the reason has not fixed the fault; it has reset the clock on it. Failure analysis is the systematic investigation of why a component failed, and its purpose is practical: a blown part is often the victim of something else, so replacing it without the cause meets the same fate (capacitor-and-inductor-failure-modes). Its central idea is that this is a process, not a guess — repeatable and evidence-based. The process follows a chain. The failure mode is how it shows — open, short, burned, cracked, drifted. The failure mechanism is the physical process that produced it — thermal runaway, dielectric breakdown, solder fatigue (building-a-troubleshooting-tree). And the root cause is the originating reason it was set in motion — overstress, thin margin, defect, wear, environment. The steps walk this chain — preserve the evidence, characterize the failure, determine the mechanism, trace to the cause, and close with a corrective action that fixes the cause, not the symptom (board-documentation-techniques). A failure is a question, and analysis is the discipline of answering why — because only the answer makes a repair permanent.

Why This Matters

This is the discipline that separates a repair that lasts from one that fails again (the-troubleshooting-process). This matters because a failed part is frequently a victim, not a culprit: a shorted MOSFET may have been killed by a downstream fault, a burned resistor by an overvoltage, a dead regulator by a load it was never meant to drive — and dropping a new part into the same conditions simply repeats the failure, so finding the cause is what makes the repair hold (capacitor-and-inductor-failure-modes). This matters because a process reaches a conclusion a guess cannot defend: an intuition names a cause that cannot be checked, while an evidence-based method traces the failure to a cause that can be shown, which is the difference between a repair a technician hopes is right and one they can stand behind (building-a-troubleshooting-tree). It matters because the chain from mode to cause is where analysis lives: the failure mode is visible and the root cause is hidden, and the whole skill is working inward through the physical mechanism to the originating reason, rather than stopping at the symptom that is merely the most obvious thing (board-documentation-techniques). And it matters because the corrective action is only as good as the cause it addresses: a repair that fixes the burned part and not the overvoltage that burned it is not a repair but a delay, so the entire purpose of the analysis is to make the corrective action land on the cause rather than the symptom. Find why the part failed, trace the chain to its root, and correct the cause — and a repair stops being a temporary reprieve and becomes a permanent fix.

Required Prerequisites

Before starting this section, you should have completed:

  • The Troubleshooting Process — the diagnostic method of finding a fault; failure analysis extends it from finding the broken part to finding why it broke.
  • Board Documentation Techniques — the as-found preservation discipline; failure analysis begins by preserving the evidence of a failure before it is disturbed.
  • A camera and the as-found habit — the failure is photographed before anything is touched, because the evidence is fragile and the first disturbance can destroy the very signature the analysis needs.
  • A failure-chain worksheet — mode, mechanism, root cause, and corrective action are written as columns, because the discipline is walking the chain deliberately, not jumping to a guess.
  • Isopropyl alcohol and swabs — a failed area is cleaned gently to read its signature clearly, but only after it is documented, since cleaning can remove evidence as easily as grime.
  • A board with a clearly failed component — a burned, shorted, or cracked part to analyze from mode through mechanism to a hypothesized cause.
  • A failure whose cause is external to the failed part — a part killed by a fault elsewhere, so the lesson that the victim is not the culprit is concrete.
  • A set of known failure signatures — burned, cracked, bulged, corroded parts to learn to read the mode and reason toward the mechanism.

Real-World Applications

Failure analysis is what a technician does to make a repair permanent instead of temporary. A repairer whose replacement part failed as fast as the first learns the lesson the whole chapter teaches — the original was a victim, and the cause is still on the board (capacitor-and-inductor-failure-modes). A technician facing a burned component preserves the evidence, reads the failure mode, reasons to the mechanism, and traces back to the cause before fitting anything (board-documentation-techniques). A bench documenting a failure for a report records the chain from mode to cause as evidence, so the conclusion can be defended rather than merely asserted (building-a-troubleshooting-tree). And a tech deciding what to actually fix aims the corrective action at the root cause — the overvoltage, the thermal path, the overload — rather than at the symptom it produced (the-troubleshooting-process). The confusions this prevents: a victim part blamed as the culprit, a guess offered where a process was needed, the symptom mistaken for the cause, and a corrective action that fixes the burned part and leaves the reason it burned.

Common Challenges

  • The symptom is loud and the cause is quiet. A burned part shouts for attention while the overvoltage that burned it leaves no mark — the analysis must look past the obvious failure to the hidden reason (board-documentation-techniques).
  • Evidence is fragile and easily destroyed. The first probe, the first cleaning, the first desoldering can erase the signature — preservation comes before investigation, or the evidence is gone (the-troubleshooting-process).
  • A guess feels faster than a process. Naming a likely cause is quicker than proving one — but an unproven cause is one the corrective action may miss, and the failure returns (building-a-troubleshooting-tree).
  • The victim looks like the culprit. The failed part is the most obvious suspect, and often the wrong one — a part is frequently killed by a fault elsewhere, and stopping at it repeats the failure (capacitor-and-inductor-failure-modes).

Safety Notes

Risk Level: Low. This section is method and reasoning — it reworks nothing of its own — but failed parts can hold hazards and the standing bench law frames it.

  • A failed part can be hazardous — bulged capacitors, sharp fractures, and burned residues are handled with the board unpowered and discharged, eyes and skin protected.
  • Powered characterization follows the live-power rules — if a live failure must be observed, it is done current-limited and watched, per the earlier chapters.
  • ESD discipline throughout — the parts around a failure are static-fragile, so the board is handled by its edges on a grounded mat.

Professional Tips Before Starting

  • Preserve before you probe. Photograph and record the failure as-found — the first disturbance can destroy the signature, and evidence lost is a cause unfindable (board-documentation-techniques).
  • Walk the chain, do not jump it. Mode to mechanism to cause, in order — skipping to a guessed cause is how the real one is missed (building-a-troubleshooting-tree).
  • Suspect the victim last. The failed part is the obvious suspect and often innocent — ask what could have killed it before blaming it (capacitor-and-inductor-failure-modes).
  • Demand evidence for the cause. A cause you can show beats one you assert — the process exists to reach a conclusion that holds (the-troubleshooting-process).
  • Aim the fix at the cause. The corrective action targets the root, not the symptom — fix the overvoltage, not just the part it burned.

The Purpose and the Process of Failure Analysis

Why Analysis, Not Just Replacement

The purpose of failure analysis is best understood against the thing it is not: the simple replacement of a broken part (the-troubleshooting-process). Replacement fixes what is visibly broken, and no more. A blown MOSFET is removed and a new one fitted, a burned resistor swapped, a dead regulator changed — and if the failed part failed on its own, of old age or a random defect, that is the whole repair. But a failed part is very often not the culprit. It is the victim of something else — a downstream short that drew the MOSFET past its limit, an overvoltage that cooked the resistor, a load fault that killed the regulator — and the thing that killed it is still there when the new part goes in, so the replacement meets the same conditions and fails the same way, sometimes in seconds. Failure analysis is the discipline that closes this gap. It is the systematic investigation of why the part failed — not merely that it did — undertaken so that the cause is found and addressed, and the repair therefore lasts rather than simply resetting the clock on the same failure (capacitor-and-inductor-failure-modes). This is the same lesson the power chapter taught about a protecting PMIC, generalized: a part that failed because of something else cannot be repaired by replacing it alone, and the whole value of analysis is finding the something else. Replacement fixes the symptom; analysis finds the cause — and the difference is whether the repair is permanent or merely postponed.

The Chain — Mode, Mechanism, and Root Cause

Failure analysis works along a chain of three linked things, and keeping them distinct is the core of the discipline (building-a-troubleshooting-tree). The first is the failure mode — how the failure shows itself. It is the observable manifestation: an open circuit, a short, a burned or vented package, a cracked solder joint, an intermittent connection, a parameter drifted out of tolerance — the thing a technician first sees or measures, the symptom in its most literal form. The second is the failure mechanism — the physical process that produced that mode. Beneath a short is a mechanism: a dielectric that broke down, a junction that ran away thermally, a metal trace that migrated, a solder joint that fatigued and cracked under thermal cycling — the actual physics of what happened to the material of the part, one layer below the visible mode. The third is the root cause — the originating reason the mechanism was set in motion. The dielectric broke down because of an overvoltage; the junction ran away because a downstream short drew excess current; the solder fatigued because of a thermal-expansion mismatch the design did not accommodate — the why behind the how, which may lie in an overstress event, a thin design margin, a manufacturing defect, the wear of age, or the environment (capacitor-and-inductor-failure-modes). The analysis moves along this chain. From the visible mode, it reasons inward to the mechanism that must have produced it, and from the mechanism back to the cause that must have set it in motion — because the mode is what you see, the mechanism is what happened, and the cause is what a repair must change. Mode, mechanism, root cause — three links a professional never collapses into one, because confusing the symptom with the cause is exactly the error the discipline exists to prevent.

The Process — Preserve, Characterize, Trace, Correct

What makes failure analysis reliable rather than lucky is that it is a process — a repeatable sequence of steps that leads anyone who follows it to a defensible conclusion (the-troubleshooting-process). The first step is to preserve the evidence. A failure's signature is fragile — a burn pattern, a fracture surface, the position of debris, a discoloration — and the first probe, cleaning, or desoldering can destroy it, so the failure is documented as-found before anything is touched, exactly the as-found discipline the documentation chapter established, because evidence lost is a cause that can no longer be found (board-documentation-techniques). The second step is to characterize the failure. The failure mode is identified precisely — what failed, and how — and its physical signature read at magnification, turning "it's broken" into a specific, described mode that the next step can reason from. The third step is to determine the mechanism and trace to the cause. From the characterized mode, the technician reasons to the physical mechanism that produced it, and from the mechanism back through the circuit and its history to the root cause — gathering evidence to confirm or reject each candidate rather than asserting one, which is where the troubleshooting discipline of forming and testing hypotheses does its work (building-a-troubleshooting-tree). The final step is the corrective action. The repair is aimed at the root cause — the overvoltage suppressed, the thermal path improved, the overload removed, the design margin restored — and only then is the failed part replaced, so that the new part meets changed conditions rather than the ones that killed the old. Preserve, characterize, trace, correct — the repeatable process that turns a failure from a mystery into an answered question, and a repair from a gamble into a certainty.

Common Mistakes

  • Replacing the part without finding the cause. The broken component is swapped and the job called done — if the part was a victim, the cause remains and the failure returns (capacitor-and-inductor-failure-modes).
  • Collapsing the chain to the symptom. The visible mode is treated as the cause — the mode is what you see, the cause is what a repair must change, and confusing them fixes nothing (building-a-troubleshooting-tree).
  • Destroying the evidence before reading it. The failure is probed, cleaned, or desoldered before it is documented — the signature is fragile, and preservation must come first (board-documentation-techniques).
  • Offering a guess in place of a process. A likely cause is named without evidence — an unproven cause may be the wrong one, and the corrective action then misses (the-troubleshooting-process).
  • Fixing the symptom and calling it enough. The burned part is replaced and the overvoltage left in place — a corrective action must change the cause, or it is only a delay.

Troubleshooting Guidance

  • A replacement part failed as fast as the original — the part was a victim: the cause is still present, so stop replacing and analyze — preserve, characterize the mode, and trace to the cause before fitting another (capacitor-and-inductor-failure-modes).
  • You can see what failed but not why — walk the chain inward: from the visible mode reason to the physical mechanism that produced it, then from the mechanism back through the circuit to the originating cause, rather than stopping at the symptom (building-a-troubleshooting-tree).
  • The evidence seems ambiguous — check that it was preserved: a signature probed or cleaned before documentation may be corrupted, so work from the as-found record and, next time, preserve before investigating (board-documentation-techniques).
  • You have a candidate cause but no proof — test it as a hypothesis: gather evidence that would confirm or reject the candidate, because a corrective action built on an unproven cause may address the wrong thing (the-troubleshooting-process).

Verification & Testing Methods

Confirm your grasp of the purpose and process before the chapter's techniques:

  • [ ] I can explain the purpose of failure analysis and why replacement without a cause is an incomplete repair.
  • [ ] I can distinguish a failure mode from a failure mechanism from a root cause, and describe the chain between them.
  • [ ] I can follow the failure-analysis process as repeatable, evidence-based steps rather than a guess.
  • [ ] I can preserve the evidence of a failure before it is disturbed.
  • [ ] I can frame a corrective action that addresses the root cause rather than the symptom.

Then try the practice exercises below — reasoning and record-building only; scenarios differ from the quiz.

Practice Exercises

  1. Preserve the failure (5 minutes, donor board with a failed part). Photograph and record a failed component as-found — its mode, its signature, its context — before touching or cleaning it, so the fragile evidence is secured before any investigation (board-documentation-techniques).
  2. Characterize the mode (5 minutes, same failure). Describe precisely what failed and how — open, short, burned, cracked, drifted — reading the physical signature at magnification, turning "it's broken" into a specific mode the next step can reason from (capacitor-and-inductor-failure-modes).
  3. Trace the chain (5 minutes, desk reasoning). From the characterized mode, reason to the physical mechanism that produced it and then back to a candidate root cause, writing the mode, mechanism, and cause as a chain rather than jumping to a guess (building-a-troubleshooting-tree).
  4. Aim the fix at the cause (5 minutes, desk reasoning). For the traced cause, state the corrective action that would address it — and why replacing the failed part alone would not — so the fix lands on the cause rather than the symptom (the-troubleshooting-process).

These core steps — the preserved failure, the characterized mode, the traced chain, and the aimed corrective action — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

  • Failure analysis is the systematic investigation of why a part failed, not just that it did — because a failed part is often the victim of something else, and replacing it without the cause resets the clock on the failure rather than ending it (capacitor-and-inductor-failure-modes).
  • A failure is traced along a chain — the failure mode is how it shows, the failure mechanism is the physical process that produced it, and the root cause is the originating reason it was set in motion — and the discipline is never collapsing these into one (building-a-troubleshooting-tree).
  • Failure analysis is a process, not a guess — repeatable and evidence-based — so it reaches a cause that can be shown and defended rather than an intuition that only sometimes holds (the-troubleshooting-process).
  • The process is preserve, characterize, trace, and correct — the evidence secured before it is disturbed, the mode described precisely, the chain traced inward to mechanism and back to cause, and the fix aimed at the root (board-documentation-techniques).
  • A corrective action addresses the root cause, not the symptom, because fixing the burned part and leaving the overvoltage that burned it is not a repair but a delay — and only correcting the cause makes the repair permanent.

Skills Learned

After completing this section, you can:

  • Explain the purpose of failure analysis and why replacement without a cause is an incomplete repair.
  • Distinguish a failure mode from a failure mechanism from a root cause.
  • Follow the failure-analysis process as repeatable, evidence-based steps.
  • Preserve the evidence of a failure before it is disturbed.
  • Frame a corrective action that addresses the root cause rather than the symptom.

Glossary Additions

New terms introduced in this section:

  • failure analysis — the systematic, evidence-based investigation of why a component or system failed, as distinct from simply identifying and replacing the broken part. Its purpose in repair is to find the cause behind a failure rather than only its symptom, because a failed part is frequently the victim of a condition elsewhere — a downstream short, an overvoltage, an overload, a thermal problem — that remains after the part is replaced and produces the same failure again. Failure analysis proceeds as a repeatable process that traces a failure along a chain from its observable mode, through the physical mechanism that produced it, to the root cause that set the mechanism in motion, and closes with a corrective action that addresses that cause. It is what makes the difference between a repair that lasts and one that merely resets the clock on the same failure.
  • failure mechanism — the physical process by which a component actually failed, the middle link of the chain between the observable failure mode and the underlying root cause. Where the failure mode is how the failure shows itself — an open, a short, a burned package, a cracked joint — the failure mechanism is what physically happened to the material of the part to produce that mode: a dielectric breaking down, a semiconductor junction running away thermally, a metal trace migrating under current, a solder joint fatiguing and cracking under repeated thermal cycling. Identifying the mechanism is the step that connects the visible symptom to its origin, because a given mode can arise from different mechanisms and each mechanism points toward different root causes, so naming the mechanism correctly is what lets the analysis trace back to the true cause rather than a plausible-sounding wrong one.
  • corrective action — the change a repair makes to address the root cause of a failure, as opposed to the replacement that addresses only its symptom. Because a failure's root cause is the condition that set its mechanism in motion — an overvoltage, an inadequate thermal path, an overload, a thin design margin — a corrective action is the step that changes that condition: suppressing the overvoltage, improving the heat path, removing the overload, restoring the margin. Replacing the failed part is necessary but not sufficient, because a new part dropped into the unchanged conditions meets the same fate; the corrective action is what makes the difference, and a failure analysis is complete only when its conclusion has been turned into an action that changes the cause rather than merely renewing the victim.

Suggested Next Sections

Must read next:

  • Root Cause Analysis Techniques — Section 5.2 takes the hardest link of the chain to its own depth: the disciplined techniques for separating the true originating cause of a failure from the symptoms and intermediate failures that mask it.

Recommended: