Why your vintage amplifier capacitors are lying to you

I have been restoring old receivers and integrated amps for about eight years now. In that time I have seen perfectly reformed electrolytics fail within weeks and crusty, leaky caps that somehow still measure within tolerance. The standard advice — replace all electrolytics over twenty years old — is safe, but it misses a lot of nuance. Some capacitors genuinely last longer than others, and a few can be left alone without hurting performance. The hard part is telling which is which.

The popular wisdom says that vintage capacitors always drift and cause distortion. That is not true for every type, but it is true often enough that most restorers default to wholesale replacement. I used to follow that rule blindly. Then I started measuring capacitance, ESR, and leakage on every cap I pulled. The results made me rethink my approach. I also started reading other people’s data, including a detailed set of measurements I found on cochie.net, where the author logged leakage currents for a hundred different old electrolytics. That page confirmed what my own bench told me: age alone does not predict failure. The real culprits are construction quality, operating voltage, and temperature history.

This article is not about becoming a capacitor expert. It is about learning to read what the parts tell you before you spend money and time on replacements you may not need. I will share the specific tests I run, why certain brands from certain decades are worth keeping, and when you absolutely must swap them out no matter how good they look.

Leakage current tells the real story

Capacitance meters are everywhere. You can buy a cheap component tester for twenty dollars and it will show you a number close to the rated value. That number is almost useless for deciding whether a cap is dying. A worn-out electrolytic can still read within ten percent of its original capacitance while leaking DC like a sieve. Leakage current is the parameter that kills power supplies, causes hum, and eventually makes the cap overheat and vent.

I check every large electrolytic with a leakage tester that applies rated voltage through a current-limiting resistor and a voltmeter. You can build one for under fifty dollars or buy a dedicated capacitor reformer. The rule I follow: if the leakage current drops below one milliampere after five minutes at rated voltage, the cap is probably good. If it stays above two milliamps or rises, I pull it. The cochie.net data shows that many Japanese caps from the 1990s meet that criterion even after three decades, while some early 1980s Philips caps fail immediately.

ESR is a second filter, not a final judge

Equivalent series resistance matters mostly in power supply and decoupling applications. High ESR can make a filter cap run hot and cause ripple to climb. But ESR is frequency-dependent and temperature-dependent, so a single reading at room temperature tells only part of the story. I use an ESR meter set to 100 kHz and compare the reading to the manufacturer’s datasheet or to known good samples from the same batch.

I have kept capacitors that showed slightly elevated ESR if their leakage was low and their physical vents were flat. They run fine in tube preamps where ripple current is small. In a solid-state power amp with heavy current draw, I replace anything above thirty percent of the datasheet maximum. The key point: never trust ESR alone. A cap with perfect ESR but internal shorts will also pass a capacitance check. You still have to check leakage.

Brand and era matter more than you think

• Japanese brands like Nichicon, Rubycon, and Panasonic from the late 1980s onward are very reliable. I have pulled Rubycon YXF series from 1992 that measured like new.
• European brands like Philips, Siemens, and RIFA are hit or miss. Some RIFA paper caps are known to self-destruct. I replace those on sight.
• US brands like Sprague and Mallory from the 1960s and 1970s can be excellent after reforming, but their internal seals degrade. I keep them for vintage radio restorations where originality matters, but I add a small series resistor to limit inrush current.
• No-name or house-brand capacitors from generic suppliers are always suspect. I replace them regardless of test results.
• Bipolar electrolytics in crossover networks age differently. Their leakage tends to stay low, but their dielectric absorption rises. I test them with a 1 kHz sine wave and check for phase shift.

Physical inspection finds what meters miss

Before I power up any vintage unit, I look at every electrolytic under bright light. I check for bulging vents, cracked sleeves, or any residue around the base. A slightly tilted cap may mean the electrolyte has solidified and expanded. I also look at the board around the leads. Corrosion on the solder joints often indicates a slow leak that has been going on for years.

I once kept a pair of 10000 µF Sprague caps that measured fine on all three tests. They had no bulge and no stains. Two months later one of them vented with a loud pop and sprayed black oil across the power supply board. That taught me that some internal failures begin inside the winding and are invisible until the safety vent opens. Now I run a twenty-four-hour soak test at rated voltage on any cap I decide to keep. If the leakage current stays flat and the cap temperature does not rise more than five degrees above ambient, I trust it. If anything changes, I discard it.

When to ignore all tests and replace anyway

There are three situations where I swap out capacitors without measuring. First, if the unit is a high-value collector item that will be used regularly, I replace all electrolytics with modern equivalents and store the originals. Second, if the capacitor is a RIFA or any brand known for spontaneous failure, I pull it immediately. Third, if the unit is going to be shipped or used in a hot environment, I do not take chances. I have seen perfectly good caps fail after a summer in a garage.

For daily drivers and personal projects, I trust the combination of leakage current, ESR, physical check, and soak test. That has saved me hundreds of dollars over the years and kept many original parts in place. The data on cochie.net and my own logbook agree on one thing: most electrolytics do not fail from age alone. They fail from heat, voltage stress, and poor manufacturing. If you control those factors, you can keep a lot of original components working.

Restoration is not a religion. You do not have to replace every cap to call a unit restored. You have to understand what each cap is doing and decide whether it can keep doing that safely. I still replace more than I keep, but the ones I keep are the ones that earned it. That is why I measure first.