Bridge Failure Signatures in Power Supplies: Burn Patterns, Shorts, and Open Arms

Updated: October 28, 2026

This page is a focused engineering note in the bridge rectifier cluster—use the checklists and links below to finish selection, not as a substitute for the datasheet.

Topic hub: Bridge rectifier guide. Related deep-dives sit in the continue list at the end.

Before release	Pass criteria
Ratings vs mission profile	Voltage/current/thermal margins recorded
Package & PCB	Rth path agreed with layout
Docs	Datasheet rev + PCN policy filed
Alternates	At least one workable second source or risk accepted

Reading the Supply: Which Symptom Should Point at the Bridge First

Not every power-supply failure starts with the bridge, and the first diagnostic habit is knowing which symptoms should point at it. A supply that produces no DC at all is more often a blown input fuse, an open transformer, or a dead switch than a dead bridge — the bridge sits downstream of those. The bridge’s failures announce themselves differently: sagging DC with normal input, ripple that changed frequency, a rail that reads fine at no load and collapses under load, or heat concentrated at the rectifier stage.

The symptom hierarchy puts the bridge in its place. Voltage absence points upstream first; voltage sag and frequency change point at the bridge and capacitor; heat and smoke point at thermal overload, which can be the bridge or the load. Reading down that hierarchy prevents the classic round of replacing parts upstream while the real failure sits in the rectifier all along.

The order also respects effort: the cheapest checks run first, and the bridge’s terminal test only happens after the upstream and capacitor causes are cleared, so a technician rarely desolders a module that a fuse or a capacitor explained. The rectifier failure-modes guide frames the same symptom-to-cause logic in engineering terms; this article adds the visual and board-level evidence the guide deliberately leaves out.

Visual Burn Signatures: Cracks, Discoloration, and Around-the-Lead Evidence

The visual inspection comes before any meter, and the burn patterns are surprisingly specific. A cracked or split diode body — usually along the glass or plastic wall — is the signature of repeated thermal cycling snapping the package. A dark burn ring around a lead is the signature of a resistive connection that ran hot, often a crimped or loose terminal rather than the junction itself. A uniformly discolored body is the signature of continuous overtemperature, while a hot spot near one corner of a four-diode module points at one arm carrying more than its share.

The burn evidence has a limit worth stating: it tells you something ran too hot, but not always what. A cracked body on one diode of a discrete bridge is clear — that diode was stressed. A burned lead on a molded module may be a terminal issue. The visual pass narrows the candidates and sets the expectations for the meter test, which is the next section’s job.

The axial package’s failure signs are covered in the axial rectifier maintenance guide, and the thermal design guide explains why thermal cycling produces the cracking pattern — the junction-to-case thermal mismatch under repeated expansion and contraction.

Always Short, Sometimes Open: Pinning the Failed Arm

The terminal checks pin the failure to a specific arm, and the pattern of “always short, sometimes open” is the signature of the two common bridge failure modes. A shorted arm reads as a conduction path in both directions: the meter shows a forward drop both ways on the affected terminal pair, because the junction has collapsed into a low-resistance path. An open arm reads OL in the forward direction, because the junction no longer conducts at all.

The pinned-arm method is the seven-terminal sequence from the bridge testing guide, run on the failed supply after it is safely powered down and discharged. The single-diode checks isolate which quarter failed; the double-diode checks confirm whether it is one arm or the whole loop. The bridge testing method is the tool; the evidence pattern here tells the technician which readings to expect from a hot, burnt, or intermittently failing part.

A half-open arm is the hardest pattern. The bridge reads healthy cold, then sags under load as the partially degraded junction warms and its resistance climbs. That pattern forces the loaded-VF verification, because a cold meter reading cannot see a failure that only appears with current and temperature.

What the Meter Can and Cannot Tell You First

Before desoldering anything, it is worth being precise about what the instrument in your hand can prove. A diode-mode reading shows that a junction conducts one way and blocks the other; it does not show the forward drop at the bridge’s rated current, and it does not show whether the thermal path is still intact.

Reading	What it establishes	What it does not
Forward drop one way, open the other, on all four pairs	The junction structure is intact	The part’s behaviour at rated current
Conduction both ways on one pair	That arm is shorted	Whether the failure is inside the die or the package
Open both ways on one pair	That arm is open	Whether it was overcurrent or thermal cycling that opened it
Reading drifts while probing	Something in the circuit is charging or discharging	Anything about the bridge until the node is discharged

The forward-drop comparison is where the numbers help. onsemi quotes the 1N400x family at 0.93 V typical and 1.1 V maximum at 1 A and 25 °C, while a meter’s diode mode injects a much smaller test current — so a healthy part reads well below that figure, and the useful comparison is between the four arms of the same bridge rather than against the datasheet column. An arm that reads noticeably higher or lower than its three siblings has failed differently from the rest.

Two boundary conditions decide whether you desolder at all. A bulk capacitor downstream still holds charge and will drive current through the bridge while you probe, so discharge it and confirm with a voltmeter first. And reverse leakage rises with temperature in the same family, from 10 µA maximum at 25 °C to 50 µA maximum at 100 °C, which is why a bridge that measures perfectly on a cold bench can still be the part that failed in a warm enclosure. The full seven-check sequence is in Bridge Rectifier Testing.

Capacitor, Bridge, or Both: Rule-Out Sequence Before Desoldering

Before the bridge is condemned, three non-bridge causes must be ruled out, and each has a fast test. First, the input stage: confirm AC actually reaches the bridge — a blown fuse or open transformer upstream produces the same “no DC” symptom. Second, the capacitor: a smoothing capacitor that has failed open reduces the rail and changes the ripple, and a swollen or leaking electrolytic is a visual red flag that needs no meter. Third, the load: a load that draws excess current heats everything and can cook the bridge that is actually fine.

The rule-out sequence runs in that order: verify input, check the capacitor, check the load, and only then read the bridge. The capacitor-bridge confusion is common because both produce sagging DC — but they leave different evidence: a failed open capacitor leaves ripple and heat at the cap, while a failed bridge arm leaves heat at the rectifier and the frequency change. The ripple and noise troubleshooting guide decodes the waveform difference, which is the fastest way to separate the two.

Field Data: Cost of a Bridge Repair vs Replacement Strategy

The last decision is economic, and the data is simple: a discrete four-diode bridge costs four small diodes and twenty minutes; a molded bridge module costs the module and a desolder. The repair-versus-replace strategy turns on the bridge type, the availability of parts, and the confidence of the diagnosis.

For a discrete bridge with one confirmed failed arm, replacing one diode is the cheap, correct call — four diodes are loose parts and the arm is isolated. For a molded module, the whole unit is replaced, because the internal junctions are not serviceable and a partial failure in a sealed package is a reliability risk even if the other three arms test fine. The cost difference between the two strategies is exactly the difference between the general rectifier family, where you buy diodes, and the standard bridge family, where you buy modules.

The failure-evidence table summarizes the reading patterns a technician meets in the field:

Evidence	Likely mode	Reading to confirm	Typical fix
No DC at all	Input upstream, not bridge	Check fuse/transformer first	Upstream repair
DC sags, ripple frequency changed	One open arm	Terminal checks on all four arms	Replace the arm/module
Heat at rectifier only	Shorted or degraded arm	Double-diode checks + loaded VF	Replace arm
Burned, cracked single body	Thermal cycle on that arm	Visual + terminal	Replace arm/module
Rail fine cold, sags under load	Half-open / degraded junction	Loaded VF at operating current	Replace suspected arm
Swollen capacitor + heat	Capacitor, maybe bridge	Cap ESR + ripple check	Replace cap first

The table is the diagnosis in one view: each row links the observed evidence to the confirming reading and the fix. A technician who works the table top-down mostly lands on the right part in one pass, and the rule-out column prevents the classic mistake of replacing the bridge when the capacitor or the input was the cause.

A worked field case ties it together. A 12 V bench supply arrives with sagging DC output and a warm rectifier stage. The input checks clear, the capacitor reads healthy, and the terminal checks show the “− to AC2” arm reading OL in the forward direction — one open arm. The visual pass on the discrete bridge shows no burn, consistent with a junction that opened quietly rather than cooked itself. The fix is one diode from the general rectifier family, and the supply goes back together with the same module. That is the bridge-failure workflow in miniature: symptom hierarchy, rule-out, pinned arm, targeted repair.

The strategy also weighs repeated failures. If the same bridge arm fails twice in the same supply, the diagnosis is wrong or the environment is hostile: either the surge protection upstream is missing, the capacitor is stressing the bridge, or the thermal path is inadequate. Before any desoldering, the safety sequence belongs in the workflow: power off the supply, disconnect the input, and discharge the filter capacitor — a charged electrolytic holds enough energy to be lethal, and it is exactly the component sitting on the same rail as the bridge. The discharge step is covered in the power-supply repair workflow context of safe servicing, and it is the one step that must never be skipped regardless of how confident the diagnosis is. The rectifier reliability checklist turns that repeat-failure observation into a pre-repair audit, so the rework does not repeat the original mistake.

FAQs

How do I tell whether the bridge or the capacitor failed?

Work with the supply off and the bulk capacitor discharged. Check each arm in diode mode and compare the four against each other, then check the capacitor for a short. A failed capacitor frequently takes the bridge with it, so both have to be verified.

Can I test a bridge rectifier in circuit?

Yes, with the power off and the capacitor discharged. Surrounding components change the readings, so the useful comparison is between the four arms of the same bridge rather than against an absolute datasheet figure.

Why does a bridge measure fine but fail under load?

A meter cannot see the thermal path or the leakage at operating temperature. A die that has lost its attachment, or a part leaking several times more at 100 degrees, passes every diode check and still fails once current flows.

Is the bridge always the cause when it is burnt?

No. A burnt bridge is often the victim rather than the cause, so check the input filtering, the inrush path and the load side before replacing it, or the replacement will fail the same way.

Release checklist

• Re-read absolute maximum ratings against worst-case spikes.
• Confirm pinout/package drawing matches the footprint library.
• Capture why this MPN beat the next alternative (loss, cost, lead time, or qualification).

FAQ

What failure signatures show up first on a bridge?

Open arms, shorted diodes, and burn patterns near the hottest die—confirm with terminal checks before swapping the whole PSU.

How do I size a bridge for a DC motor drive?

Use stall/current peaks and LR dynamics, not only nameplate RMS; verify ripple current into the DC bus.

Can I breadboard a bridge from four 1N4007s?

Yes for learning and low-power checks; production still needs rated I_FSM, creepage, and thermal path.

Basics?

What is a bridge rectifier.

Continue through this engineering topic

• Bridge rectifier guide
• Common Cathode Dual Diodes vs Two Discretes When the Ct
• Fast Recovery Bridge Rectifier Guide High Frequency Input Rectificatio
• Half Wave vs Full Wave Rectification Reading the Waveform Instead
• Schottky bridge rectifier guide low forward drop for low voltage ac inputs
• Single phase vs three phase bridge rectifiers size for the input not the output

Topic hub: Bridge rectifier guide — related selection guides and application notes in this cluster.