Capacitor Failure Symptoms

Capacitors provide filtering, decoupling, timing and coupling in every piece of vintage equipment on this wiki, and they are the parts most likely to have changed value or failed outright since it left the factory. This page covers what failure looks like, how to confirm it with a meter, and what the manufacturers actually say about replacement — as opposed to what gets repeated on forums.

Not every capacitor family fails the same way. An aluminium electrolytic dries out and its ESR climbs; a tantalum goes short and can catch fire; a Class II ceramic loses most of its capacitance under DC bias without ever looking wrong; a mains-rated paper suppression capacitor cracks its case and fills the room with smoke. Diagnosis starts with knowing which kind you are looking at.

Common symptoms of capacitor failure

Symptom	What is happening	Where it turns up
Bulging or domed top	Gas generated inside the can has lifted the pressure relief vent. Nichicon lists the causes as reverse voltage, over-voltage, extreme ripple or AC voltage on a polarised part; end-of-life electrolyte evaporation produces the same effect more slowly	Power supplies, motherboards, monitor analogue boards
Leaking electrolyte	The seal has failed and electrolyte has crept out onto the board, where it attacks copper, solder and neighbouring parts	Surface-mount electrolytics on late-1980s and 1990s boards, analogue boards, audio gear
No power, or intermittent start-up	Reservoir or filter capacitance has fallen far enough that the rail collapses under load, or the supply's start-up circuit no longer works	Televisions, amplifiers, consoles, PCs
Whine, buzz or chirp from a switching supply	High ESR in the output filtering pushes the converter out of its intended operating region; some supplies audibly hiccup instead of starting	Power bricks, monitors, all-in-one machines
Rippling, shrinking or flickering picture	Failed capacitors in the video, deflection or B+ supply	Monitors and televisions
Hum or buzz in audio	Failed smoothing capacitance in a linear supply. The dominant tone is twice the mains frequency after a full-wave rectifier — 100 Hz on a 50 Hz supply, 120 Hz on 60 Hz — so a 50 Hz buzz points at something other than the reservoir capacitor	Receivers, amplifiers, synthesisers
Regulators or power ICs running hot	Ripple the filtering is no longer removing is being dissipated downstream	Anything with a linear regulator after a rectifier
Resets, crashes and memory errors	Rail noise from failed decoupling, most often on 5 V logic	Computers, routers, set-top boxes

What fails, by capacitor family

Aluminium electrolytic

A vented electrolytic capacitor

The dominant part on vintage boards and the one with a genuine wear-out mechanism: the liquid electrolyte escapes through the seal over time, capacitance falls and ESR rises. Nichicon divides failure into catastrophic (the capacitor has completely lost its function through a short or an open circuit) and degradation (gradual deterioration against the limits in the part's own specification). In the field it lists short circuit — which it describes as very rare, and caused by vibration, shock, lead stress, over-voltage, extreme ripple or pulse current — open circuit, degradation of capacitance and ESR, and operation of the pressure relief vent.

Visual signs:

• Bulging or split vent on the top, or a domed base on a surface-mount part
• Fluid or dried crust at the base, and discoloured laminate around it
• Corroded or blackened leads, and green or dark residue on nearby pads
• A cracked or lifted sleeve

A capacitor that looks perfect can still be dead. Axial and older radial parts frequently dry out with no external sign at all.

Tantalum

Solid tantalum capacitors have no known wear-out mechanism, according to Vishay — but excessive voltage, current or temperature reduces their reliability, and the failure is a short circuit that can lead to overheating and possibly ignition. On a vintage board this is the part that goes bang, takes the fuse or the regulator with it, and leaves a charred hole.

• Usually fails short
• Smoke, a flash or a small fire; sometimes a crater
• Blackened, cracked or split case
• Frequently no warning at all beforehand

Derating matters when you replace one. Vishay's guidance is normally 60 % to 50 % voltage derating for solid tantalum parts. A 6.3 V tantalum on a 5 V rail — which is what a great many 1980s machines shipped with — is running at about 80 % of rating, and that is a substantial part of why they fail. Fitting 16 V or 20 V parts in those positions is standard practice in restoration. Vishay also notes that a fuse in series changes the failure mode from a short to an open, which is why some designs have one.

Ceramic

Ceramic capacitors have no electrolyte and no wear-out mechanism worth worrying about, but they fail in two ways that are easy to miss:

• Mechanical cracking. Board flex, a dropped machine, or rework heat cracks the body. The part may then short — often intermittently, and often only when warm.
• Class II dielectric behaviour. X7R, Y5V and similar high-permittivity types lose capacitance with applied DC bias, with temperature, and slowly with age. Nichicon's own comparison of dielectrics puts barium titanate — the Class II ceramic dielectric — at a permittivity of 500 to 20,000, against 15 to 250 for the Class I temperature-compensating types; that enormous permittivity is exactly what makes it bias- and temperature-dependent. A "0.1 µF" decoupling capacitor measuring well under its marked value on a meter at low bias is often behaving as designed, not faulty.

A cracked ceramic gives no outward sign in most cases. It is found by measurement, or by finding a rail pulled down.

Film, and mains-rated suppression capacitors

Plastic film capacitors (polyester, polypropylene, polystyrene) are generally the most stable parts on an old board and rarely need replacing on principle. The exception is the mains-rated suppression capacitor sitting across the line or from line to earth in the power supply input, classified X or Y under EN/IEC 60384-14.

These parts are built to fail safely rather than to never fail. KEMET's PME271M — the metallised impregnated paper Class X2 type used across the line in an enormous number of 1970s and 1980s supplies — is encapsulated in self-extinguishing material meeting UL 94 V-0 and is qualified against IEC 60384-14 for both active and passive flammability, including a needle-flame test. The flammability qualification exists because the designers expect some of them to fail.

In restoration work the characteristic end for these parts is a loud crack, a burst case and a sharp, persistent smell of fish, with smoke pouring from the supply. This is widely reported by restorers and is one of the few "replace it before you power the machine on" recommendations on this wiki; it is an observation from the restoration community rather than a published manufacturer figure, and no manufacturer failure rate for the aged parts is quoted here because none has been found. Replace across-the-line capacitors with correctly class-rated parts — X2 for across the line, Y2 for line to earth — and never with an ordinary film capacitor of the same value and voltage.

Testing a suspect capacitor

Method	Tool	What it tells you, and what it does not
Visual inspection	None	Finds vented, leaking and cracked parts. Finds nothing else
ESR measurement	ESR meter	The single most useful test on an electrolytic, and it works in circuit on most boards because the test signal is small. Compare against the part's own datasheet figure or against an identical good part; an absolute reading with no reference is close to meaningless
Capacitance measurement	Capacitance or LCR meter	Lift one leg. In circuit you are measuring everything in parallel with the part. Standard tolerance on an aluminium electrolytic is ±20 %
Leakage	Bench supply with current limit, or a meter with a leakage function	Charge through a resistor to rated voltage and watch the current settle. Nichicon notes that leakage current rises during long unpowered storage and comes back down when voltage is reapplied
Continuity / resistance	Multimeter	Finds hard shorts, which on a vintage board usually means a tantalum. Will not find a dried-out electrolytic
Scope on the rail	Oscilloscope	Ripple on a DC rail under load is the direct evidence that filtering has failed. This is the test that confirms a diagnosis rather than suggesting one

Discharge before you measure. Nichicon's own handling guidance requires a capacitor to be discharged before it is removed. In a monitor or an all-in-one machine, see CRT Discharge Procedure first.

Replacing capacitors

• Voltage rating: match or exceed the original. A rating at or below the rail the part sits on is always wrong, and it damages machines.
• 105 °C parts in power supplies. Life follows the Arrhenius relationship — Nichicon states that life doubles for every 10 °C drop in temperature. A 105 °C part in place of an 85 °C part of the same endurance rating is a real, calculable gain, not a superstition.
• Match ESR and ripple rating, not just µF and volts. A switch-mode supply designed around a low-ESR part will misbehave with a general-purpose one. Exceeding the ripple rating is one of Nichicon's listed causes of capacitance loss and rising ESR.
• No datasheet, no purchase. Buy parts whose ESR, ripple current and endurance hours you can look up for the exact part number you are ordering. That is the test — not the brand name on the sleeve, which is the thing counterfeiters copy. See Battery Explosion, Capacitor or Corrosion Damage for the fuller discussion of choosing parts and suppliers.
• Recap by function, not by the whole board. Replace the group that is failing — the supply's secondary side, the analogue board, the surface-mount electrolytics — rather than every capacitor on every board as a reflex. Each unnecessary desoldering operation is another chance to lift a pad.

Cleaning up after a leak

• Remove the failed parts first — the residue keeps coming out of the can otherwise.
• Wash the affected area with warm water and a little washing-up liquid, working it with a soft brush, then rinse thoroughly with clean water.
• Dry completely — compressed air, then warmth. Water under a socket or a connector will cause trouble later.
• Finish with isopropyl alcohol to lift flux and any remaining film. Nichicon warns that non-ionic halides left behind by some fluxes migrate into capacitors and corrode them, so clean flux off properly rather than leaving it.
• Inspect every track and via in the affected area under magnification before powering up. Electrolyte travels along traces, under parts and through vias to the other side of the board.

Things that are not faults

• A decoupling ceramic reading low. Class II dielectrics lose capacitance under DC bias and with age by design.
• Capacitance 15 % off the marked value on an electrolytic. Standard tolerance is ±20 %.
• A slightly domed base on a surface-mount electrolytic. Check for residue and measure it; the case shape alone is not conclusive.
• A warm capacitor in a power supply. Ripple current produces heat. Hot enough to be uncomfortable to touch is a problem; warm is normal.
• A bipolar or non-polar electrolytic with no stripe. It is not a mismarked part. Nichicon notes that polarised types cannot be used in continuous AC applications, which is what the non-polar ones are for.

Storage, and the truth about "reforming"

The folk advice to "power up vintage gear now and then to reform the capacitors" has a real procedure underneath it, and the real procedure is more specific.

Nichicon states that an aluminium electrolytic stored unpowered for a long period develops a slightly raised leakage current, because the anode oxide layer reacts with the electrolyte. Applying voltage re-forms the oxide and the leakage current returns to its initial level. The formal treatment for a loose capacitor is to charge it to its rated voltage through a resistance of about 1 kΩ and hold it there for about 30 minutes. For a capacitor already in a circuit, set the supply to roughly half the rated voltage for about ten minutes, then raise it gradually while watching the equipment; if the supply is not adjustable, run it for about thirty minutes and then switch off.

Crucially, Nichicon also states that a capacitor stored for less than two years at 5 °C to 35 °C can be used without voltage treatment. So:

• Re-forming is a real procedure with real numbers, and it applies to leakage current, not to lost capacitance or high ESR. Nothing reverses a dried-out capacitor.
• Occasionally switching a machine on is not a re-forming procedure and does not follow the guidance above.
• Storage conditions are worth getting right: Nichicon recommends 5 °C to 35 °C and 75 % relative humidity or less, out of direct sunlight.

Related pages

• Battery Explosion, Capacitor or Corrosion Damage
• Category:Capacitor Replacement Guides — machine-specific replacement lists
• CRT Discharge Procedure
• Recommended Tools