The Repair LibraryRead · Learn · Master

PMIC Failure Analysis

Section 3.2 taught the PMIC as an architecture to read; this section, at Professional depth, is what a technician does when that architecture fails — and its whole discipline turns on one distinction that separates the professional from the parts-swapper. A dead rail is not a diagnosis. It is a symptom with several possible true causes, and the most important thing a failure analysis must do is decide which one it is before anything is replaced. The rail may be dead because the PMIC's own regulator for it has genuinely failed. It may be dead because a short downstream is dragging it to ground and the PMIC, entirely healthy, has folded that rail back or shut it down to save itself — a rail short tripping the chip's overcurrent protection and dropping it into a protection mode that looks exactly like death. It may be dead because an enable was never asserted or a sequence upstream never completed, so the rail was never commanded to exist. Or it may be gated off by a load switch that never closed. These are different faults with one appearance, and the section's method is the rail-by-rail discipline that tells them apart: work the architecture in order — input, always-on, enables, each rail's presence and voltage — and for every rail that is dead or low, ask not merely whether it is there but why it is not, measuring its resistance to ground, its current, its heat, to separate a rail the PMIC failed to make from a rail something else is dragging down. The stakes of getting this wrong are concrete and expensive. A PMIC that shut a rail down to protect itself from a downstream short is a healthy chip, and a technician who reads its protection as failure and replaces it has done two harmful things at once: thrown away a good part, and installed a new one straight onto the same short, which will trip its protection or destroy it in turn. So the section's creed is find the true cause first — clear the short, restore the enable, close the switch — and never replace a PMIC that was only doing its job.

ProfessionalMedium Risk23 min read

What You Will Learn

  • You will learn that a dead rail is a symptom with several true causes, and that failure analysis is deciding which before replacing.
  • You will learn the rail-by-rail method — working the architecture in order and asking not just whether each rail is present but why not.
  • You will learn to tell a genuine PMIC failure from a protection mode the healthy chip entered in response to a fault.
  • You will learn the rail short — a downstream short to ground that trips overcurrent protection and makes a healthy rail look dead.
  • You will learn why finding the true cause must precede replacement, and how replacing a protecting PMIC destroys its successor.

What You Will Be Able To Do

  • You will be able to treat a dead rail as a symptom and enumerate its possible true causes before acting.
  • You will be able to work a PMIC rail by rail against its architecture, asking why each dead rail is not present.
  • You will be able to distinguish a PMIC in protection mode from a genuinely failed PMIC.
  • You will be able to find a rail short by resistance, current, and heat, and locate it with the short-hunting methods.
  • You will be able to clear the true cause before replacing a PMIC, and explain why replacing a protecting chip destroys the new one.

Required Tools

  • A multimeter with a low-ohms range — to measure a rail's resistance to ground and find a rail short
  • A bench supply with current display — to watch a rail's current draw and see a PMIC enter protection
  • A thermal camera or freeze spray — to find the hot component a downstream short is running, referencing the short-localization sections
  • The notebook — this section records each rail's state, cause, and the protection-versus-failure verdict

When NOT to Attempt This

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

  • You are not experienced with the specific repair type described here.
  • You do not have professional-grade equipment for this procedure.
  • The device has sentimental or high monetary value and you cannot afford a mistake.
  • You have not successfully completed this repair on a sacrificial device first.

Section Overview

Section 3.2 taught the PMIC as an architecture to read; this closes on diagnosing it when it fails, rail by rail, at Professional depth (pmic-architecture-in-modern-devices). A dead rail is not a diagnosis — it is a symptom with several true causes. The PMIC's own regulator failed, a downstream short is dragging the rail down, an enable never fired, or a load switch never closed (switching-power-supply-theory). The method is rail-by-rail discipline. Work the architecture in order — input, always-on, enables, each rail's presence and voltage — and for every dead rail ask not just whether it is there but why it is not. The heart of it is one distinction. A rail short downstream trips the chip's overcurrent protection and drops it into a protection mode that looks exactly like death — but the PMIC is healthy and reacting (diagnosing-regulator-and-converter-faults). Reading protection as failure is the costly error. Replace a protecting PMIC and you throw away a good part and install its successor onto the same short (isolating-the-shorted-component). So find the true cause first — clear the short, restore the enable, close the switch — and never replace a chip that was only doing its job.

Why This Matters

This is the difference between a professional power repair and an expensive guess (pmic-architecture-in-modern-devices). This matters because a dead rail's appearance hides its cause: the same missing voltage can be a failed regulator, a downstream short, a missing enable, or an open switch, and these demand opposite actions — so a technician who reads the symptom as the diagnosis acts blindly, while one who enumerates the causes first acts correctly (switching-power-supply-theory). This matters because protection mode is a trap for the parts-swapper: a PMIC that shut a rail down to survive a short is a healthy chip that looks dead, and replacing it is the single most common expensive mistake in power repair — a good part discarded and a new one killed on the fault that was never found (diagnosing-regulator-and-converter-faults). It matters because the true cause is often not at the PMIC at all: a rail short lives in a downstream capacitor or load, so the fix is there, not in the chip that merely reported it by shutting down — and finding it takes the short-hunting methods, not a reflow of the PMIC (isolating-the-shorted-component). And it matters because the method scales to any device: rail-by-rail against the architecture is a discipline, not a memorized fix, so it works on a phone, a laptop, or a console equally, turning any failed power system into an ordered investigation. Enumerate the causes, work rail by rail, separate protection from failure, and find the true cause before replacing anything — and PMIC failure analysis becomes a method a technician can trust.

Required Prerequisites

Before starting this section, you should have completed:

  • PMIC Architecture in Modern Devices — the architecture this section diagnoses; a rail-by-rail analysis is impossible without the map of rails, enables, and load switches it provides.
  • Switching Power Supply Theory — the regulated rail whose failure modes this section analyzes, and the working model that says what a healthy rail should look like.
  • Isopropyl alcohol and freeze spray — a rail short's hot component is found by warmth or by freeze spray, and clean contacts make low-ohms readings trustworthy.
  • Note cards for a rail table — each rail's expected voltage, measured voltage, resistance to ground, and verdict is written in a table, because rail-by-rail analysis is a record, not a memory.
  • Fine flux and fresh solder — a confirmed failed component is removed and replaced cleanly, because a rework that starves a joint on a power rail creates the next fault.
  • A donor board with a known dead rail — a real failed power system to work rail by rail, ideally one whose cause is a downstream short rather than the PMIC.
  • A board with a deliberately shorted rail capacitor — a made rail short, so the protection-versus-failure distinction can be practiced safely and the short located end to end.
  • A bench supply with a current display and a thermal camera — to watch a rail draw excess current into a short and to see the hot component the short is running.

Real-World Applications

This method is what separates a fixed power fault from a repeated one. A technician facing a phone with a dead core rail works the architecture in order, finds the rail pulled to a low resistance to ground, and diagnoses a downstream short rather than a failed PMIC (isolating-the-shorted-component). A repairer whose replacement PMIC died as fast as the first learns the lesson of protection mode — the original chip was healthy and protecting against a short that the new chip was installed straight onto (diagnosing-regulator-and-converter-faults). A bench diagnosing a rail that is simply absent, not shorted checks its enable and sequence before condemning the regulator, because a rail never commanded on is not a failed rail (pmic-architecture-in-modern-devices). And a tech confirming a genuinely failed PMIC has first cleared every downstream cause — no rail short, enables present, switches closed — so the chip is replaced once, correctly, onto a board that will not kill it (switching-power-supply-theory). The confusions this prevents: a dead rail treated as a diagnosis, a protecting PMIC swapped as failed, a downstream short reflowed at the chip, and a replacement installed onto the fault that killed the first.

Common Challenges

  • The symptom looks the same for every cause. A dead rail reads the same whether the PMIC failed, a short pulls it down, or an enable never fired — only asking why, rail by rail, separates them (switching-power-supply-theory).
  • Protection perfectly imitates failure. A healthy PMIC in protection mode is dark and dead-looking — distinguishing it from a failed chip is the section's hardest and most valuable skill (diagnosing-regulator-and-converter-faults).
  • The true cause hides downstream. A rail short lives in a load or a capacitor away from the PMIC — the chip only reported it, and the fix is at the short, found by the short-hunting methods (isolating-the-shorted-component).
  • Replacement is tempting and destructive. Swapping the PMIC feels like progress — done before the cause is cleared, it destroys the new chip on the old fault and teaches nothing (pmic-architecture-in-modern-devices).

Safety Notes

Risk Level: Medium. This section probes and injects on live power rails and reworks confirmed-failed parts, so the standing bench law, the always-on caution, and the heat discipline all apply.

  • Current-limited injection only — hunt a short with a low, current-limited voltage and a thermal camera, never by forcing high current, which can heat a shorted part to damage or injury.
  • Disconnect for resistance and rework — the always-on rail means resistance readings and physical work are done with the battery and input out.
  • Clear the short before re-powering — a confirmed rail short is cleared before any PMIC is powered again or replaced, so the fault does not claim the next chip.

Professional Tips Before Starting

  • Write the rail table first. Expected voltage, measured, resistance to ground, verdict — the analysis is a record, and the table is what keeps it honest (pmic-architecture-in-modern-devices).
  • Ask why, not just whether. A dead rail's cause matters more than its absence — enumerate the causes before you reach for a part (switching-power-supply-theory).
  • Suspect protection before failure. A dead rail with a downstream short is a healthy PMIC protecting itself — rule out the short before condemning the chip (diagnosing-regulator-and-converter-faults).
  • Find the short with the right methods. Resistance, current, and heat localize it — use the short-hunting techniques, not a reflow of the PMIC (isolating-the-shorted-component).
  • Clear the cause before you replace. A new PMIC on an old short dies too — the true cause is cleared first, always.

Analyzing a Failed Power System

The Dead Rail Is a Symptom — Enumerate the Causes

The whole of PMIC failure analysis rests on refusing to accept a dead rail as a diagnosis (switching-power-supply-theory). A missing rail voltage is a symptom, and a symptom has causes — several of them, all with the same appearance. The rail may be dead because the PMIC's own regulator for it has failed: an internal switch, a burned bond, a genuinely dead converter inside the chip. It may be dead because a downstream short is dragging it to ground: a shorted decoupling capacitor, a failed load, a via bridged to a plane, pulling the rail low no matter how hard the PMIC tries. It may be dead because it was never commanded on: a missing enable, a sequence that stalled, a control-bus configuration never written, so the rail is absent not because it failed but because nothing told it to exist. Or it may be gated off: a load switch that never closed, leaving the rail present at the PMIC and absent past the switch, exactly as Section 3.2 warned. These four causes demand four different actions, and they cannot be told apart by the symptom alone, because the symptom — a rail that should be some voltage and is zero — is identical for all of them. This is why the first move is enumeration, not action. Before a meter probes deeply or a part is touched, the technician names the possible causes and resolves to distinguish them, because a repair that treats the symptom as the diagnosis fixes the wrong thing or nothing. A dead rail is a question, not an answer — and the analysis is the disciplined work of answering which cause is real.

The Rail-by-Rail Method — Working the Architecture in Order

The way to answer that question is a method: work the PMIC's architecture in order, and for every rail ask not just whether it is present but why it is not (pmic-architecture-in-modern-devices). The order follows the hierarchy Section 3.2 built. Start at the input — is the PMIC even powered — then the always-on rail, then the enables and the sequence that should bring the main rails up, then each main rail in turn, and finally the load-switched branches beyond them, because a fault early in that chain explains everything downstream of it and checking in order keeps the search from chasing symptoms of a single upstream cause. For each dead or low rail, the questions are specific. What voltage should it be, and what is it? What is its resistance to ground — is it a healthy finite value or the low ohms of a rail short? What current is the rail or the whole board drawing — normal, or the elevated draw of something dragging it down? Is anything getting hot? These measurements separate the causes. A rail low with a near-zero resistance to ground is being shorted, not failed by the PMIC; a rail simply absent with a normal resistance and no enable is a sequence or command fault; a rail the PMIC drives but that reads wrong with nothing dragging it is a genuine regulator fault (diagnosing-regulator-and-converter-faults). The rail table is the tool that holds it together. Every rail's expected voltage, measured voltage, resistance, and current go in a table, and the pattern across the table — one rail shorted, or all rails absent, or one regulator wrong — points at the cause far better than any single reading. Input, always-on, enables, each rail, each branch, every one asked why — the ordered method that turns a dead power system into a located fault.

Protection Is Not Failure — The Distinction That Saves the Chip

The single most valuable judgment in the whole section is telling a healthy PMIC that is protecting itself from one that has actually failed (diagnosing-regulator-and-converter-faults). Every modern PMIC defends itself. It watches each rail for excess current, overvoltage, and its own temperature, and when a fault threatens it, its overcurrent protection — along with thermal and overvoltage shutdowns — folds the rail back or shuts it down to survive, dropping the chip into a protection mode that keeps it alive while the fault persists. And a PMIC in protection mode looks exactly like a dead one. The rail is off, the chip is dark, the device will not start — the very picture of a failed PMIC — except that the chip is completely healthy and is reacting correctly to a fault that is somewhere else, most often a rail short downstream pulling its output to ground. Reading this protection as failure is the costliest error in power repair. A technician who condemns the protecting chip and replaces it does two harmful things at once: discards a perfectly good part, and installs the new one directly onto the same short that tripped the first, so the replacement immediately enters its own protection or, if the short is hard enough, is destroyed outright (isolating-the-shorted-component). The distinction is made by looking for the fault, not the chip. A dead rail with a low resistance to ground and elevated current is a PMIC protecting against a short; a dead rail with normal resistance and correct enables, that the PMIC is genuinely failing to produce, is a candidate for a failed chip — and only after every downstream cause is cleared is the PMIC itself judged (switching-power-supply-theory). Protection wears the mask of failure — and the technician who looks behind it saves the chip, finds the real fault, and fixes the device once.

Common Mistakes

  • Accepting the dead rail as the diagnosis. The missing voltage is treated as the fault — it is a symptom with several causes, and acting before enumerating them fixes the wrong thing (switching-power-supply-theory).
  • Swapping a protecting PMIC as failed. A healthy protecting chip is read as failed and replaced — the good part is wasted and the new one is installed onto the same short (diagnosing-regulator-and-converter-faults).
  • Reflowing the PMIC for a downstream short. The chip that reported the fault is reworked — the rail short lives downstream and is found with the short-hunting methods, not at the PMIC (isolating-the-shorted-component).
  • Skipping the enable and sequence. An absent rail is blamed on the regulator — a rail never commanded on by an enable or sequence never failed, and the check is fast (pmic-architecture-in-modern-devices).
  • Replacing before clearing the cause. The PMIC is changed with the fault still present — the new chip meets the same short, and the repair repeats.

Troubleshooting Guidance

  • A rail is dead and you do not know why — enumerate and measure: name the four causes — failed regulator, downstream short, missing enable, open switch — then measure the rail's resistance to ground, its current, and its heat to tell them apart, because the symptom is identical across all four (switching-power-supply-theory).
  • A rail is low with a near-zero resistance to ground — a rail short, not a failed PMIC: the rail is being dragged down, so the PMIC is likely protecting itself, and the short is located downstream with the short-hunting methods before the chip is touched (isolating-the-shorted-component).
  • A replacement PMIC failed as fast as the first — protection mode missed: the original was healthy and protecting against a fault still on the board, so stop replacing, find and clear the true cause, and only then fit a chip (diagnosing-regulator-and-converter-faults).
  • A rail is simply absent with normal resistance — check the enable and sequence: a rail never commanded on is not a failed rail, so confirm its enable and the sequence that should bring it up before condemning the regulator (pmic-architecture-in-modern-devices).

Verification & Testing Methods

Confirm your failure-analysis skill before moving to the chapter's final section:

  • [ ] I can treat a dead rail as a symptom and enumerate its true causes — failed regulator, downstream short, missing enable, open switch — before acting.
  • [ ] I can work a PMIC rail by rail against its architecture, asking why each dead rail is not present, and record it in a rail table.
  • [ ] I can distinguish a PMIC in protection mode from a genuinely failed PMIC.
  • [ ] I can recognize a rail short by low resistance to ground and elevated current, and locate it with the short-hunting methods.
  • [ ] I can explain how overcurrent protection makes a healthy PMIC look dead, and why clearing the true cause must precede replacement.

Then try the practice exercises below — rail-table analysis and short-safe diagnosis, under the callout's power, injection, and heat discipline.

Practice Exercises

  1. Build the rail table (6 minutes, donor board with a dead rail, unpowered then measured). For a board with a dead rail, list every rail with its expected voltage, then measure each rail's voltage and its resistance to ground, filling a table, so the analysis begins as an ordered record rather than a guess (pmic-architecture-in-modern-devices).
  2. Work rail by rail (7 minutes, same board). Following the architecture's order — input, always-on, enables, main rails, load-switched branches — reason from the table which rails are shorted, which are absent, and which read wrong, isolating the level at which the fault begins (switching-power-supply-theory).
  3. Separate protection from failure (7 minutes, board with a made short). On a board with a deliberately shorted rail, confirm the rail's low resistance and elevated current, and state whether the PMIC is protecting or failed — then locate the short with a low current-limited injection and a thermal camera, practicing the distinction that saves the chip (isolating-the-shorted-component).
  4. Clear before replacing (5 minutes, desk reasoning from the analysis). For the analyzed board, write the true cause, the fix, and the explicit order of operations — clear the cause first, verify, and only then judge or replace the PMIC — so the discipline of never fitting a chip onto an uncleared fault is fixed as a rule (diagnosing-regulator-and-converter-faults).

These core steps — the rail table, the rail-by-rail reasoning, the protection-versus-failure verdict, and the cause cleared before replacement — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

  • A dead rail is a symptom, not a diagnosis — its true cause may be a failed regulator, a downstream short, a missing enable, or an open load switch, and failure analysis is deciding which before anything is replaced (switching-power-supply-theory).
  • The method is rail-by-rail against the architecture — input, always-on, enables, each rail, each branch — asking of every dead rail not just whether it is present but why it is not, in a rail table that holds the pattern (pmic-architecture-in-modern-devices).
  • A rail short downstream drags a rail to a low resistance to ground and elevated current, so a rail low with near-zero resistance is being shorted, not failed by the PMIC — and the short is found downstream with the short-hunting methods (isolating-the-shorted-component).
  • A PMIC's overcurrent protection and its thermal and overvoltage shutdowns drop it into a protection mode that looks exactly like failure — a healthy chip, dark and dead-looking, reacting correctly to a fault elsewhere (diagnosing-regulator-and-converter-faults).
  • Reading protection as failure wastes a good chip and destroys its replacement on the same short — so the true cause is always cleared first, and only a PMIC proven failed with every downstream cause ruled out is replaced.

Skills Learned

After completing this section, you can:

  • Treat a dead rail as a symptom and enumerate its possible true causes before acting.
  • Work a PMIC rail by rail against its architecture, asking why each dead rail is not present.
  • Distinguish a PMIC in protection mode from a genuinely failed PMIC.
  • Find a rail short by resistance, current, and heat, and locate it with the short-hunting methods.
  • Clear the true cause before replacing a PMIC, and explain why replacing a protecting chip destroys the new one.

Glossary Additions

New terms introduced in this section:

  • rail short — a short circuit from a power rail to ground — or to another rail — somewhere downstream of the regulator that produces it, most often a failed decoupling capacitor, a shorted load, or a bridged via, which drags the rail toward zero volts no matter how hard the regulator drives it. Its defining signature is a low resistance from the rail to ground — far below the healthy finite value a good rail shows — together with elevated current draw as the regulator pushes into the short. Its importance in failure analysis is that it makes a healthy regulator look dead: the rail short pulls the rail down and trips the PMIC's protection, so the missing rail is a symptom of the short and not of a failed chip, and the short must be located downstream and cleared — using the resistance, current, and thermal short-hunting methods — before the PMIC is judged or replaced.
  • overcurrent protection — the mechanism by which a PMIC limits or shuts off a rail that is drawing more current than it should, protecting both the chip and the board from a fault such as a downstream short. When a rail's current exceeds its limit, overcurrent protection responds by folding the rail back to a reduced level, or by latching it off, or by retrying it periodically in a hiccup pattern, so that the regulator survives a fault it cannot supply. Its significance to a technician is that it is a healthy, deliberate response, not a failure: a PMIC whose overcurrent protection has shut a rail down is working exactly as designed, so the presence of overcurrent protection means a dead rail must be checked for the excess current that would trip it before the chip itself is suspected.
  • protection mode — the state a healthy PMIC enters when one of its protections — overcurrent, overvoltage, undervoltage, or thermal — has acted to shut down or fold back a rail in response to a fault, keeping the chip alive while the fault persists. A PMIC in protection mode is dark and dead-looking: the rail is off, the device will not start, and every outward sign matches a failed chip, which is exactly why the mode is a trap. The distinction that matters is that the chip is healthy and reacting correctly to a fault elsewhere — typically a downstream rail short — so a PMIC in protection mode is not replaced but investigated: the fault that put it into protection is found and cleared, after which the chip resumes normal operation, and only if it does not is it judged genuinely failed.

Suggested Next Sections

Must read next:

  • Power Sequencing Failures — Section 3.4 closes the chapter on the timing-order faults this analysis sets up: the cases where every rail is present yet the device will not start because the rails came up in the wrong order or at the wrong time, and the method for reading a sequence that has gone wrong.

Recommended: